Radical ions of crownophanes derived from tetraphenylethene. Solution structures and ion-pairing phenomena

Article abstract of DOI:10.1021/jo980920o

Radical anions and radical cations of parent tetraphenylethene (1) and its cyclophane derivatives 2-8 were generated by chemical methods. These paramagnetic species were characterized by their ESR and simultaneously recorded optical spectra. The hyperfine coupling constants were substantiated by ENDOR and their signs by general TRIPLE spectroscopy. The structures of the radical ions were established with the help of quantum chemical calculations. The electron distribution in the phane radical ions is closely related to that found for 1^.-/1^.+. In the radical anions no specific complexation of alkali-metal counterions, K⁺ and Li⁺, by the crown-ether moieties could be observed. When the radical cations were generated by oxidation with Tl(III) trifluoroacetate in 1,1,1,3,3,3-hexafluoropropanol, distinctly different ESR spectra were recorded as with the use of other oxidants. This observation is rationalized by the characteristic solvation properties of 1,1,1,3,3,3-hexafluoropropanol.

Full text of DOI:10.1021/jo980920o

8
806
J . Org. Chem. 1998, 63, 8806-8814
Ra d ica l Ion s of Cr ow n op h a n es Der ived fr om Tetr a p h en yleth en e.
Solu tion Str u ctu r es a n d Ion -P a ir in g P h en om en a
†
‡
,‡
§
Fr e´ d e´ rique Barbosa, Vincent P e´ ron, Georg Gescheidt,* and Alois F u¨ rstner
Institute of Organic Chemistry, Department of Chemistry, University of Basel, St.-J ohanns-Ring 19,
CH-4056 Basel, Institute of Physical Chemistry, Department of Chemistry, University of Basel,
Klingelbergstrasse 80, CH-4056 Basel, and Max-Planck-Institut f u¨ r Kohlenforschung,
Kaiser Wilhelm-Platz 1, D-45470 M u¨ lheim/ Ruhr
Received May 13, 1998
Radical anions and radical cations of parent tetraphenylethene (1) and its cyclophane derivatives
2
-8 were generated by chemical methods. These paramagnetic species were characterized by their
ESR and simultaneously recorded optical spectra. The hyperfine coupling constants were
substantiated by ENDOR and their signs by general TRIPLE spectroscopy. The structures of the
radical ions were established with the help of quantum chemical calculations. The electron
•
-
•+
distribution in the phane radical ions is closely related to that found for 1 /1 . In the radical
+
+
anions no specific complexation of alkali-metal counterions, K and Li , by the crown-ether moieties
III
could be observed. When the radical cations were generated by oxidation with Tl trifluoroacetate
in 1,1,1,3,3,3-hexafluoropropanol, distinctly different ESR spectra were recorded as with the use
of other oxidants. This observation is rationalized by the characteristic solvation properties of
1
,1,1,3,3,3-hexafluoropropanol.
In tr od u ction
Organic molecules consisting of numerous π centers
often are efficient electron donors and acceptors. There-
fore, such organic molecules have been regarded as
1
-11
constituents of, e.g., redox switches
or of organic
wires.1
2-17
F igu r e 1. Frontier orbitals of 1 according to H u¨ ckel calcula-
tions.
The oxidation and the reduction potentials are closely
related to the orbital energies of the frontier orbitals. For
a π system, the energy gap between the highest occupied
and lowest unoccupied orbital (HOMO-LUMO gap)
in this respect is tetraphenylethene (1) consisting of 26
conjugated π centers.
2
decreases when the number of sp -hybridized centers is
increased. Consequently extended π systems can readily
accept and donate electrons. One of the early examples
†
Institute of Organic Chemistry.
Institute of Physical Chemistry.
Max-Planck-Institute f u¨ r Kohlenforschung.
‡
§
(
1) Daub, J .; Salbeck, J .; Knoechel, T.; Fischer, C.; Kunkely, H.;
Rapp, K. M. Angew. Chem. 1989, 101, 1541.
2) Daub, J .; Fischer, C.; Salbeck, J .; Ulrich, K. Adv. Mater. 1990,
, 366.
3) Daub, J .; Fischer, C.; Gierisch, S.; Sixt, J . Mol. Cryst. Liq. Cryst.
Sci. Technol., Sect. A 1992, 177.
4) Daub, J .; Beck, M.; Knorr, A.; Spreitzer, H. Pure Appl. Chem.
996, 68, 1399.
5) Achatz, J .; Fischer, C.; Salbeck, J .; Daub, J . Chem. Commun.
991, 504.
(
Both radical anions and radical cations of 1 generated
by one-electron reduction and oxidation reactions have
2
(
been analyzed in terms of their hyperfine data obtained
(
18-20
from their ESR spectra.
pling constants, a
The proton-hyperfine cou-
1
1
•
-
•+
H
, of 1 and 1 indicate identical
(
patterns due to the alternant connectivity of the π
(
(
(
(
6) Bross, P. A.; Schoeberl, U.; Daub, J . Adv. Mater. 1991, 3, 198.
7) Knorr, A.; Daub, J . Angew. Chem., Int. Ed. Engl. 1995, 34, 2664.
8) Spreitzer, H.; Daub, J . Liebigs Ann. 1995, 1637.
21
centers. The HOMO and the LUMO of 1 according to
the H u¨ ckel model are shown in Figure 1.
9) Scholz, M.; Gescheidt, G.; Schoeberl, U.; Daub, J . Magn. Reson.
In the oxidized and in the reduced stage, the central
ethylene double bond is weakened because, for the radical
cation, an electron is removed from an orbital with a
strongly bonding interaction between the ethene C atoms,
whereas for the radical anion, the interaction between
Chem. 1995, 94.
(
(
(
(
10) Tsivgoulis, G. M.; Lehn, J . M. Adv. Mater. 1997, 9, 39.
11) Tsivgoulis, G. M.; Lehn, J . M. Adv. Mater. 1997, 9, 627.
12) Baumgarten, M.; Muellen, K. Top. Curr. Chem. 1994, 169, 1.
13) Bohnen, A.; Koch, K. H.; Luettke, W.; Muellen, K. Angew. Chem.
1
990, 102, 548.
14) Klaerner, G.; Mueller, M.; Morgenroth, F.; Wehmeier, M.;
Soczka, G. T.; Muellen, K. Synth. Met. 1997, 84, 297.
15) Morgenroth, F.; Reuther, E.; Muellen, K. Angew. Chem., Int.
Ed. Engl. 1997, 36, 631.
16) Mueller, M.; Morgenroth, F.; Scherf, U.; Soczka, G. T.; Klaerner,
G.; Muellen, K. Philos. Trans. R. Soc. London, Ser. A 1997, 355, 715.
17) Quante, H.; Schlichting, P.; Rohr, U.; Geerts, Y.; Muellen, K.
Macromol. Chem. Phys. 1996, 197, 4029.
(
(18) Tabner, B. J . J . Chem. Soc. A 1969, 2487.
(19) Lundgren, B.; Levin, G.; Claesson, S.; Szwarc, M. J . Am. Chem.
Soc. 1975, 97, 262.
(20) Levin, G.; Claesson, S.; Szwarc, M. J . Am. Chem. Soc. 1972,
94, 8672.
(21) Heilbronner, E.; Bock, H. The HMO-Model and its Application;
Wiley and Verlag Chemie: London, New York, and Weinheim, 1976.
(
(
(
1
0.1021/jo980920o CCC: $15.00 © 1998 American Chemical Society
Published on Web 11/10/1998

Solution Structures and Ion-Pairing Phenomena
J . Org. Chem., Vol. 63, No. 24, 1998 8807
Sch em e 1. Rea ctivity of 1•+ Accor d in g to Ref 25
•+
mechanism leading to the occurrence of 3 , however, is
unknown yet.
In this paper, we report on the generation of the radical
anions and the radical cations of tetraphenylethene
derivatives 4-8. In these molecules two 1,2 phenyl
groups are connected by a polymethylene (4) or polyoxy-
ethylene (5-8) bridge either in the para (4-7) or in the
ortho (8) positions.
these centers becomes antibonding. Therefore, isomer-
izations of tetraphenylethene-type molecules should occur
preferably in these open-shell stages.
The geometrical consequences of an electron transfer
to or from 1 were also inspected by electrochemical
methods2
2,23
and electronic-absorption spectroscopy.24
H
Five sets of hyperfine coupling constants, a , each due
•
-
to four equivalent protons were reported for 1 (0.209
The altered electronic structure, geometry, and the
chelating ability of the oxygen atoms in the (crowno)-
phane moieties of 4-8 in respect to the parent compound
1 should lead to distinct changes of the geometry and
the reactivity of the corresponding radical ions. The
mT (4 H, para), 0.167 mT (4 H, ortho), 0.136 mT (4 H,
ortho′), 0.081 (4 H, meta), and 0.036 mT (4 H, meta′)).18
These values are almost matched by those ascribed to
•
+
1
(0.201, 0.150, 0.090, 0.083, and 0.050 mT, respec-
tively).2 This shows that the phenyl rings do not rotate
5
•
-
•+
H
hitherto published a values of parent 1 /1 were only
freely on the hyperfine time scale and that the electron
distribution in the radical anion and the radical cation
are rather similar, in line with the pairing properties
suggested by the H u¨ ckel model.
derived from the ESR spectra, and their signs were not
yet determined experimentally; therefore, we have also
reexamined these radical ions experimentally by the
2
6
ENDOR and TRIPLE techniques and by quantum
The conspicuous electron-accepting properties of 1 are
chemical calculations.27
indicated by the observation that the dianion 12 is
-
2
3
formed at almost the same potential as the monoion
and often occurs as a byproduct when 1•- is generated
by alkali-metal reductions of 1 in ethereal solvents. This
Exp er im en ta l Section
The radical anions of 1-8 were generated by dissolving the
neutral compounds in 1,2-dimethoxyethane (DME) or tetrahy-
drofuran (THF) and reduced on a K mirror under high vacuum
•
-
is corroborated by the diminished ESR intensity of 1
when the formation of tight ion pairs is promoted (high
temperature, nonpolar solvents); the additional increase
of the ESR line widths indicates the shorter lifetime of
-
5
(10 Torr). The solvents were dried by refluxing over Na/K
alloy for 4 days followed by distillation into a Schlenk tube
containing Na/K alloy. After degassing, the solvents were
•
-
•-
2-
1
by a shift in the equilibrium 21 h 1 + 1 .
-
5
•+
stored under high vacuum (10 Torr).
It has been shown that 1 undergoes further reactions
2 2
The oxidation reactions were performed in CH Cl (dried
(
Scheme 1).25 Indeed the ESR spectrum attributable to
the follow-up product dibenzophenanthrene 3 grew into
over molecular sieves under vacuum), trifluoroacetic acid
TFA), or 1,1,1,3,3,3-hexafluoropropan-2-ol (HFP) (all solvents
•
+
(
•+
the ESR signal of 1 with gradually increasing intensity
within ca. 0.5 h; this could also be followed by additional
lines in the ENDOR spectrum). Presumably the forma-
highest available purity, Merck). Solutions of the oxidants
III
(
(Tl trifluoroacetate, tris(p-bromophenyl)ammoniumyl hexachlo-
roantimonate), and the parent compounds were mixed at low
-
5
tion of 3 proceeds via 1,2-diphenyl derivative 2.25 The
•
+
temperature (193 K) under vacuum (10 Torr). The reactors
were transferred to the cavity of the ESR instrument, and the
temperature was increased until an ESR signal occurred.
(22) Grzeszczuk, M.; Smith, D. E. J . Electroanal. Chem. Interfacial
Electrochem. 1984, 162, 189.
(
23) Shultz, D. A.; Fox, M. A. J . Org. Chem. 1990, 55, 1047.
(26) Kurreck, H.; Kirste, B.; Lubitz, W. Electron Nuclear Double
Resonance Spectroscopy of Radicals in Solution; VCH: Weinheim,
1988.
(27) Batra, R.; Giese, B.; Spichty, M.; Gescheidt, G.; Houk, K. N. J .
Phys. Chem. 1996, 100, 18371.
(24) Suzuki, H.; Koyano, K.; Shida, T.; Kira, A. Bull. Chem. Soc.
J pn. 1979, 52, 2794.
25) Eberson, L.; Hartshorn, M. P.; Persson, O. J . Chem. Soc., Perkin
Trans. 2 1995, 1735.
(

8
808 J . Org. Chem., Vol. 63, No. 24, 1998
Barbosa et al.
Ta ble 1. Selected a H of Ra d ica l An ion s 1•- a n d 4•--8•- (All Given a H Ar e Attr ibu ted to Tw o Equ iva len t P r oton s Excep t
•
-
for 8 (1H))
aH/mT (for the labeling of the positions, see the formulas)
T/K meta meta′ meta meta′ ortho ortho′ ortho ortho′ para
1•
-
-
-
-
-
-
-
-
-
-
-
/ K / THF 273 +0.036 +0.081 +0.036 +0.081 -0.136 -0.167 -0.136 -0.167 -0.209
+
•
+
4
4
5
5
6
6
6
7
/K / THF 193 +0.008 +0.022 +0.036 +0.080 -0.127 -0.146 -0.184 -0.184 -0.230 +0.252
•
+
/ Li / THF 273 +0.006 +0.037 +0.080 +0.096 -0.127 -0.149 -0.180 -0.180 -0.231 +0.251
•
+
/ K / THF 193 +0.006 +0.035 +0.035 +0.083 -0.105 -0.131 -0.161 -0.209 -0.267
•
+
/ Li / THF 193 +0.007 +0.035 +0.035 +0.084 -0.104 -0.131 -0.161 -0.209 -0.267
•
+
/ K / THF 193 +0.019 +0.036 +0.036 +0.084 -0.114 -0.142 -0.156 -0.198 -0.254
•
+
/ Na / THF 193 +0.018 +0.036 +0.036 +0.085 -0.117 -0.142 -0.156 -0.196 -0.250
•
+
/ Li / THF 193 +0.018 +0.037 +0.077 +0.086 -0.110 -0.137 -0.159 -0.203 -0.263
•
+
/ K / THF 193 -0.022 +0.042 +0.052 +0.068 -0.104 -0.124 -0.210 -0.229 -0.278
•
+
8
/K / THF
193 +0.004 +0.028 +0.043 +0.061 +0.075 +0.094 -0.106 -0.122 -0.132 -0.170 -0.185 -0.204 0.217 0.239
•
+
8
/ Li / THF 193 +0.022 +0.035 +0.046 +0.061 +0.081 +0.091 -0.111 -0.122 -0.134 -0.157 -0.185 -0.228 -0.242 -0.285
The electronic absorption spectra of the solutions containing
the radical ions were taken inside the microwave cavity in the
same region of the sample tube and simultaneously with the
ESR spectra using a specially developed coupling of fiber optics
to the cavity of the ESR spectrometer28 and a J &M (Aalen,
Germany) TIDAS diode array spectrometer (220-1021 nm).
ESR spectra were taken on a Varian E9 or Bruker ESP 300
spectrometer. The latter instrument was also used for EN-
DOR and general TRIPLE measurements.
The ESR spectral simulations were performed with the
2
9
freeware program Winsim.
The H u¨ ckel MO calculations were done with the program
3
0
31
MacHMO.
semiempirical AM1 and MNDO
initio and density functional theory (DFT) calculations were
The MOPAC 6.0 package was used for the
3
2
33,34
method whereas the ab
F igu r e 2. ESR (experimental and simulation) and ENDOR
3
5
performed with Gaussian 94.
The syntheses of 4-8 were already reported.36
•
-
+
spectra of 4 (solvent, THF; counterion, K ; temperature 243
K).
crownophane 6 should bind to K+ (and likewise Na+).
Reduction of 6 in THF on a Na mirror led to the detection
of the same aH as with K as the counterion. Thus, no
Resu lts a n d Discu ssion
+
Sim u lta n eou s in Situ Detected ESR a n d Op tica l
Sp ectr a of th e Ra d ica l An ion s. The radical anions of
and 4-8 were generated by reduction on a potassium
specific interaction between the crown-ether moiety and
these alkali-metal cations could be established.
1
1
8
mirror in THF, and their ESR spectra were studied
between 193 and 273 K. In this temperature range, the
ESR spectra remained almost unchanged and were only
poorly resolved. The application of ENDOR spectroscopy
In agreement with formerly published data, our
ENDOR spectra bear out that the parent radical anion
•
-
1
possesses five sets of four equivalent protons. As
expected, the free rotation of the four phenyl groups is
restricted and the two ortho and the two meta positions
are not equivalent. The biggest aH (0.209 mT) is assigned
to the para hydrogen whereas the aH of 0.167 and 0.136
mT stem from the two different ortho protons. The
smallest aH of 0.081 and 0.036 mT consequently are
attributed to the meta protons. A detailed assignment
of the aH can be achieved by comparing the experimental
values with the calculated ones (see below).
1
reveals that the isotropic H-hyperfine coupling con-
stants, a
). Addition of LiCl to the solutions of the radical anions
generally did not lead to the detection of significantly
H
, remain almost temperature invariable (Table
1
•
-
different ESR/ENDOR spectra; only in the case of 6
a
slight increase of the biggest a was noted. Possibly,
H
+
interaction of the crown ether moiety with Li leads to a
small change of the twist of the phenyl groups without
changing the overall C
crown-6 is a favored ligand for the K cation. Thus,
s
symmetry. The macrocycle 18-
The connection of the 1,2 diphenyl groups by a alkyl
or polyoxyethylene bridge as in 4-8 diminishes the C2v
symmetry of the parent tetraphenylethene. In 4-8 the
+
(
(
28) Gescheidt, G. Rev. Sci. Instr. 1994, 65, 2145.
29) Duling, D. R. PEST Winsim; NIEHS, Research Triangle Park,
s
maximum achievable symmetry is C with pairwise
equivalent phenyl groups. The decrease of symmetry
depends on the conformation of the connecting chain. It
NC, 1995.
(
(
30) Huber, H. MacHMO; University of Basel, Basel, 1985.
31) Seiler, F. J . MOPAC 6.0; Res. Lab., U. S. Air Force Academy,
Colorado Springs, 1994.
32) Dewar, M. J . S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J . J . P.
J . Am. Chem. Soc. 1985, 107, 3902.
•
-
•-
is shown in Table 1 that 4 -7 possess eight or nine
different a . Moreover, the ESR simulations indicate
that all a correspond to at least two equivalent protons
Figure 2) revealing that the phenyl groups are pairwise
equivalent (C symmetry). In principle, the sizes of the
have a very similar pattern as those of parent 1 . The
H
(
H
(
(
(
33) Thiel, W. QCPE No. 438; QCPE, Bloomington, IN, 1982.
34) Dewar, M. J . S. J . Am. Chem. Soc. 1977, 99, 4899.
35) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
(
s
•-
J ohnson, B. G.; Robb, M. A.; Cheeseman, J . R.; Keith, T.; Petersson,
G. A.; Montgomery, J . A.; Raghavachari, K.; Al-Laham, M. A.;
Zakrzewski, V. G.; Ortiz, J . V.; Foresman, J . B.; Peng, C. Y.; Ayala, P.
Y.; Chen, W.; Wong, M. W.; Andres, J . L.; Replogle, E. S.; Gomperts,
R.; Martin, R. L.; Fox, D. J .; Binkley, J . S.; Defrees, D. J .; Baker, J .;
Stewart, J . P.; Head-Gordon, M.; Gonzalez, C.; Pople, J . A. Gaussian
4, Revision B.2; Gaussian, Inc.: Pittsburgh, PA, 1995.
36) Fuerstner, A.; Seidel, G.; Kopiske, C.; Krueger, C.; Mynott, R.
Liebigs Ann. 1996, 5, 655.
a
H
•-
positive sign of the biggest a
that this a can be attributed to the â-methylene protons
of the decamethylene chain. Accordingly the a of 0.235
mT is assigned to the para protons of the unsubstituted
phenyl groups. The remaining a then belong to the
ortho (negative sign) and meta (positive sign) positions
H
of 4 (0.252 mT) suggests
H
H
9
H
(

Solution Structures and Ion-Pairing Phenomena
J . Org. Chem., Vol. 63, No. 24, 1998 8809
experimental spectra already reported. The use of the
ENDOR technique reveals, in agreement with Eberson’s
data, that five different a
H
values, each belonging to four
equivalent protons, lead to the observed line pattern. This
corroborates the finding that the pairs of ortho and meta
•
+
protons in 1 are not equivalent in the temperature
range from 193 to 273 K.
Oxidation of 4-8 with Tl trifluoroacetate or tris(p-
bromophenyl)ammoniumyl hexachloroantimonate in HFP,
III
2 2
trifluoroacetic acid, or CH Cl as the solvent led to the
detection of ESR spectra. In most cases the ESR signals
are attributable to the radical cations of 4-8 but often
also represent superimposed signals.
F igu r e 3. ESR (experimental and simulation) and ENDOR
spectra of 8 (solvent, DME; counterion, K ; temperature 193
K).
•-
+
III
When crownophane 5 is oxidized with Tl trifluoro-
2 2
acetate in CH Cl a partly resolved ESR spectrum is
detected. The simulation of the ESR curve with the
coupling constants taken from the corresponding ENDOR
•
-
•-
(
Table 1). In 5 -7 the highest a
H
have a negative sign
and in analogy to 1 belong to the para protons of the
,2 phenyl groups. The remaining a correspond to those
of 4 and are assigned in an analogous way.
The number of a detected in the ENDOR spectra of
8•-
is 14 vs 8 or 9 in the remaining crownophanes (Figure
, right), and the simulation of the ESR spectrum is
•
-
1
H
spectrum leads to a favorable agreement when each a
H
•
-
is attributed to two equivalent protons (Figure 5).
III
H
2 2
Oxidation of 4 with Tl trifluoroacetate in CH Cl , at
193 K leads to an ESR spectrum with a large positive aH
of +0.493 mT (2 H), marking a slow-exchange between
the two â-methylene protons on the hyperfine time scale.
At 273 K this coupling is replaced by aH ) 0.292 mT (fast
exchange, four equivalent protons). Owing to the unre-
solved ESR signal, a specific simulation of the dynamic
process of the conformational interconversion of the
decamethylene chain is not practicable. The free en-
thalpy of the interconversion can be estimated to ca. 25
3
successfully achieved when each of these coupling con-
stants is attributed to one single proton (Figure 3). This
exhibits that bridging at the ortho positions of 8 causes
a rigid arrangement of the 1,2-phenyl groups; i.e., the
free rotation of these phenyl groups is hindered and a
lowering of the symmetry arises (on the time scale of the
ESR experiment).
-1
It is noteworthy that the a
H
of the radical anions and
kJ .mol based on the difference of the coupling constants
in the slow-exchange time domain and the temperature
their multiplicities are almost independent of the tem-
perature and the counterion. No significant changes can
be detected on going from the relatively huge and
polarizable counterion K to the hard and small Li
Table 1). Comparison of the a
3
8
at which the equilibration of the two a values occurs.
H
•
+
•+
The aH of 5 and 6 represented by the ENDOR
spectrum are rather similar to those of 4 except that
+
+
•+
•
-
(
H
of 4 containing a
the aH of the â-protons are missing because the p-
methylene group is replaced by O atoms (Table 2). The
carbonyl groups in 7 with their electron-accepting effect
shift the positive charge and the spin to the two unsub-
H
•
-
decamethylene bridge with 5 in which the 10-center
bridge contains four oxygen atoms reveals only marginal
•
-
differences caused by two additional a
from two â-protons of the decamethylene group and some
deviations in the a values which can be traced back to
differing electronic effects of the CH groups and O atoms.
It seems justified to assume that the Coulombic interac-
tion between the negatively charged tetraphenylethene
π system and the alkali-metal cation dominates the
complexation at the crownophane moiety. Even in the
case of 7 where the para benzo ester group leads to a
concentration of the negative charge at the carbonyl O
H
in 4 stemming
stituted phenyl groups. Therefore the a attributed to
•
+
•+
•+
H
the para protons increase relative to 1 and 4 -6 . In
•
+
2
8
where the -O((CH ) O) (CH ) O- bridge is attached
at the ortho positions where the MO coefficients are
2
2
2
2 2
smaller than in the para positions, the a become close
H
•
+
to those of parent 1 (Table 2).
Generally, the ESR spectra ascribed to the radical
III
cations of 4-8 generated by oxidation with Tl trifluo-
roacetate in CH Cl or tris(p-bromophenyl)ammoniumyl
2
2
+
atoms, no particular ion-pairing effects by the Li cation
hexachloroantimonate can be simulated using the aH and
assuming pairwise equivalent protons. Therefore it
can be established.
The electronic absorption spectra of 1•- and 4 -8
•-
•-
seems justified that the maximum achievable C sym-
2
•
+
•+
show comparable patterns with prominent bands at ca.
00 nm and 600-800 nm (Figure 4) and agree well with
metry of 4 -8 exists on the hyperfine time-scale, i.e.,
the bridging in the para or ortho positions does not
severely affect the symmetry of the electron distribution
in the tetraphenylethene π system.
5
•
-
19,24
the spectra of 1 published previously.
Even in the
•
-
+
case of 7 at which a Li cation could also bind to the
carbonyl oxygen, addition of LiCl does not affect the
electronic spectrum (see insert in Figure 4). The band
at 600-800 nm is assigned to the SOMO-LUMO + 1
transition.
•+
•+
•+
2 2
The solutions of 1 and 4 -8 in CH Cl give rise to
almost matching absorption spectra (Figure 6). As for
the radical anions they agree very well to the spectra
24
reported by Suzuki et al. Two main bands are located
Sim u lta n eou s in Situ Detected ESR a n d Op tica l
Sp ectr a of th e Ra d ica l Ca tion s. The radical cation
of 1 was described many years before by Lewis and
Singer37 and more recently by Eberson and co-workers.
Our ESR spectra obtained after chemical oxidation of 1
at ca. 500 and ca. 880 nm and a third band around 630
nm is discernible. According to PPP calculations the
absorption at ca. 500 nm was assigned to the SOMO-
LUMO transition whereas long wave band was ascribed
to a SOMO-1-SOMO transition. It is noteworthy that
25
•+
by various methods and ascribed to 1 correspond to the
(
38) Friebolin, H. Basic One- and Two-Dimensional NMR Spectros-
(37) Lewis, I. C.; Singer, L. S. J . Chem. Phys. 1965, 43, 2712.
copy, 2nd ed.; VCH: Weinheim, 1993.

8
810 J . Org. Chem., Vol. 63, No. 24, 1998
Barbosa et al.
F igu r e 4. Electronic absorption spectra of 1 and 4•--8•- measured simultaneously with the ESR spectra (solvent, THF;
•-
+
counterion, K ; temperature 243 K).
geometry of the radical cations of 5-8 in which the spin
is unevenly distributed between the four benzene rings.
Such a distortion can be caused by an altered (rigid)
arrangement of the polyoxyethylene chains induced by
the binding of a cation to the crown ether moiety.
3
+
Obviously, with the use of Tl(III) trifluoroacetate Tl
cations are present in the solutions containing the radical
cations. It is known that HFP preferentially solvates
anions. If the trifluoroacetate anions are solvated by
HFP, “naked” Tl3 ions remain in solution and can bind
to the crown-ether moieties of 5-8. The interaction of
the radical cation with Tl3 leads to four positive charges
within the molecular assembly. However, the Coulombic
repulsion between the delocalized π radical cation and
+
+
F igu r e 5. ESR (experimental and simulation) and ENDOR
spectra of 5 (solvent, CH Cl ; counterion, CF COO ; tem-
2 2 3
perature 273 K).
•
+
-
3
+
Tl embedded in the crown ether moiety is weak because
of the remote sites carrying the positive charges.
the spectra are almost identical for all radical cations and
3+
To test whether Tl
cations, in the presence of
•
+
correspond to those of 1 generated by γ irradiation in
-
CF
spectra of solutions of 12-crown-4 and 15-crown-5 in HFP
and in CH Cl both in the presence and in the absence
3
COO bind to crown ethers in HFP, we recorded NMR
a Freon matrix.2
4
Effect of HF P a s th e Solven t. It is peculiar that
the use of HFP as solvent leads to markedly altered ESR
and ENDOR spectra particularly for 4-6 and 8 but not
for the ester 7. As an example, the ESR spectra recorded
after oxidation of 7 is shown in Figure 7.
2
2
of Tl(III) trifluoroacetate. The conspicuous observation
was that line broadening of the signal attributed to the
crown ether protons occurred only when HFP was the
solvent and the Tl(III) salt was present. Therefore it
The prominent differences between the spectra ob-
3+
seems justified to assume that Tl interacts with the
III
tained after Tl trifluoroacetate oxidation in CH
2
Cl
2
and
•+
•+
polyoxymethylene bridges in 5 -8 . It is also likely that
this effect is preferably manifested by the ESR technique
because of its faster dynamic time scale compared to
NMR. A more detailed and extended study is, however,
necessary to rigorously establish these assumptions.
in HFP as the solvent are: (i) In HFP a higher number
of coupling constants is detected. (ii) The a measured
in HFP reach significantly higher values than in CH Cl
A rationale for these observations could be the in-
H
2
2
.
•
+
tramolecular coupling reaction analogous to 1 shown
in Scheme 1.25 But none of the a
detected are consistent
H
Ca lcu la tion s
with rearrangement products such as radical cations
derived from dibenzoterphenyl (3) or 9,9′-bifluorene (9).
The method of choice for the calculation of a
H
in radical
ions of extended (planar) π systems has been the H u¨ ckel
model.2
1,39
Indeed, application of this simplistic proce-
dure in these particular cases leads to a pertinent
agreement of the experimental and the calculated values.
A prerequisite is the use of empirical perturbation
parameters which mimic the geometrical and the elec-
(
39) Gerson, F. Hochaufl o¨ sende ESR-Spektroskopie; Verlag Che-
More likely, the experimental data point to a perturbed
mie: Weinheim, 1967.

Solution Structures and Ion-Pairing Phenomena
J . Org. Chem., Vol. 63, No. 24, 1998 8811
tronic influences of the alkylidene or polyoxyethylene
bridges. The steric congestion between the facing ortho
H atoms is accounted by kortho ) -0.1â, the inductive
effect of the alkyl or O-alkyl groups by k ) -0.3, and the
O atoms in the carbonyl groups of extended π system of
7
by k
O
) 2.0 and hCO ) 1.56â.
In the recent time it was shown that density functional
theory (DFT)4
0,41
is a favorable tool for the calculation of
isotropic hyperfine coupling constants in a variety of
radicals (see, e.g. ref 27 and 42-48). In view of the
considerable size of compounds 5-8, we have first
inspected if it is necessary to optimize the geometry of
the entire molecules or if it is sufficient to regard only
the π system in which the spin and the charge are
delocalized in the radical ions. Molecular models indicate
that the polyoxyethene and the polymethylene chains
connecting the two phenyl groups possess numerous
minimum-energy conformations and allow an almost
unrestricted rotation of the phenyl groups (except 8).
•-
•+
Optimizations (UHF/3-21G*) of 7 and 7 show that the
tetraphenylethene fragment adopts an almost matching
geometry as if the crown ether moiety was replaced by a
COOH substituent. Therefore we have replaced the
chains by OCH
3
(5, 6, 8) and COOH (7) groups to reduce
the CPU times. On the basis of the minimized geometry,
single-point calculations with the B3LYP4
functional were performed. Roughly, the a
9,50
hybrid
H
calculated
by these procedures correspond to the shapes of the
H u¨ ckel frontier orbitals; however, the agreement between
the theoretical and the experimental data is best for the
ab initio/DFT combination.27 The data are compared in
Figures 8 and 9 for the radical anions and cations,
respectively, of parent 1. Analogous correlations also
hold for 5/6, 7, and 8. This is shown in Tables 3 and 4
for the radical anions and radical cations, respectively.
Para′′, para′′′ di(methoxy) substituted 1 served as model
for 5/6 whereas for 7 both the parent compound and the
model with para′′, para′′′ di(COOH) was compared.
Moreover, Table 3 contains calculations of three ion pairs
of 7 (MNDO geometry optimization, see below). The
geometry was optimized with AM1 or UHF/3-21G*
whereas the single-point calculations were performed
with the UB3LYP/6-31G*. Both protocols point to com-
parable shapes of the singly occupied orbitals. Typically,
the UHF/3-21G* geometry leads to a better agreement
of experimental and calculated data than AM1.
Geom etr y of th e Ra d ica l Ion s
The geometry of the radical ions of tetraphenylethene
has already been subject of investigations. On the basis
(
40) Kohn, W.; Becke, A. D.; Parr, A. G. J . Phys. Chem. 1996, 100,
2974.
41) Parr, R. G.; Yang, W. Density Functional Theory of Atoms and
Molecules; Oxford University Press: Oxford, 1989.
42) Adamo, C.; Barone, V.; Fortunelli, A. J . Phys. Chem. 1994, 98,
648.
1
(
(
8
(
(
43) Barone, V. Theor. Chim. Acta 1995, 3, 113.
44) Eriksson, L. A.; Malkin, V. G.; Malkina, O. L.; Salahub, D. R.
J . Chem. Phys 1993, 99, 9756.
45) Eriksson, L. A.; Malkina, O. L.; Malkin, V. G.; Salahub, D. R.
J . Chem. Phys 1994, 100, 5066.
46) Himo, F.; Eriksson, L. A. J . Chem. Soc., Faraday Trans. 1995,
1, 4343.
47) Malkin, V. G.; Malkina, O. L.; Eriksson, L. A.; Salahub, D. R.
(
(
9
(
In Theoretical and Computational Chemistry; Seminario, J . M.,
Politzer, P., Eds.; Elsevier: Amsterdam, 1995; Vol. 2.
(48) Gano, J . E.; J acob, E. J .; Sekher, P.; Subramaniam, G.;
Eriksson, L. A.; Lenoir, D. J . Org. Chem. 1996, 61, 6739.
(
(
49) Becke, A. D. Phys. Rev. A 1988, 38, 3098.
50) Becke, A. D. J . Chem. Phys. 1993, 98, 1372.

8
812 J . Org. Chem., Vol. 63, No. 24, 1998
Barbosa et al.
F igu r e 6. Electronic absorption spectra of 1•+ and 4•+-7•+ measured simultaneously with the ESR spectra (solvent, CH
2
2
Cl ;
-
counterion, SbCl
6
; temperature 263 K).
•
+
F igu r e 7. ESR and ENDOR spectra of 7 (solvent, HFP;
-
3
counterion, CF COO ; temperature 273 K).
F igu r e 9. Comparison of experimental and calculated a
H
of
•
+
1
.
electron removal or electron addition (Figure 10). As a
consequence, a twist of this bond is predicted by the AM1
model and the ab initio procedure for both radical ions.
The twist angle is rather similar in both radical ions not
exceeding 38°. For both methods, this angle is 4-8°
larger for the radical anion. Compared with the X-ray
structures of neutral crownophanes in which the corre-
sponding twist amounts to 8-14° 36 the electron-transfer
induces an additional bending of 20-30° (Figure 10).
The phenyl groups are twisted out of the ethylenic
F igu r e 8. Comparison of experimental and calculated a
H
of
•-
1
.
•-
•+
plane only by ca. 33° in 1 and 1 . For the neutral
crownophanes 5-7, these angles amount to about 50°
•-
of the optical spectroscopy, it was postulated that 1 has
•
+
36
a different geometry than 1 . Whereas in the radical
anion the central bond is reported to be twisted by a
considerable amount (up to almost 90°), it was assumed
to be almost unaffected in the radical cation.2 On the
other hand, electrochemical studies indicated that the
twist around the central bond is much less pronounced
according to the X-ray-determined geometry.
The good agreement between the experimental and the
calculated a
H
suggests that the geometry of the radical
4
ions of 1 and phanes 4-8 is very similar. This is also
borne out by AM1 and Hartree-Fock optimizations of
the radical cation and anion of 7 and 8 and model system
of 5/6, and 7 in which the twist angles of the ethenic bond
and the phenyl groups reach almost the same values as
(
e30°) and that phenyl-ring torsion plays an important
•
- 23
role for the geometry (and ion-pairing properties) of 1 .
In this respect it is interesting to compare the geometry
calculated for the neutral 1, its radical anion, and the
radical cation. In agreement to π electron models the
ethenic bond of tetraphenylethene (1) is weakened upon
•
-
•+
in 1 /1 (Figure 10).
To test the significance of these statements we have
twisted the central bond in the optimized geometry of
•
-
•+
1 /1 by 90° and computed the a
H
by the unrestricted

Solution Structures and Ion-Pairing Phenomena
J . Org. Chem., Vol. 63, No. 24, 1998 8813
F igu r e 10. Selected geometry parameters of tetraphenylethene and phane radical cations and radical anions according to quantum
mechanical calculations.
Ta ble 3. Ca lcu la ted a H of th e Ar om a tic P r oton s of th e Ra d ica l An ion s of 1, 7, a n d 8, Mod el System s of 5-7, a n d Ion
•
-
+
P a ir s of 7 /Li (for m od el system s of 5/6, see F igu r e 10)
aH/mT (for the labeling of the positions, see the formulas)
meta′ meta′′ meta′′′ othro ortho′ ortho′′ ortho′′′ para
UB3LYP/6-31G*//UHF/3-21G* +0.051 +0.102 +0.051 +0.102 -0.164 -0.192 -0.192 -0.164 -0.242
UB3LYP/6-31G*//AM1
UB3LYP/6-31G* (SinglePoint) +0.114 +0.079 +0.114 +0.079 -0.153 -0.131 -0.131 -0.153 -0.173
UB3LYP/6-31G*//UHF/3-21G* +0.068 +0.105 +0.065 +0.104 -0.183 -0.208 -0.144 -0.202 -0.275
method
meta
para′
1•
-
-
-
•
1
1
+0.078 +0.123 +0.078 +0.123 -0.184 -0.227 -0.227 -0.184 -0.272
•
90°
•
•
-
-
model of [5/6]
model of [5/6]
UB3LYP/6-31G*//AM1
+0.083 +0.124 +0.094 +0.137 -0.195 -0.231 -0.176 -0.266 -0.283
•
•
-
-
7
7
UB3LYP/6-31G*//UHF/3-21G* +0.050 +0.084 +0.105 +0.065 -0.121 -0.137 -0.136 -0.162 -0.160
UB3LYP/6-31G*//MNDO +0.053 +0.089 +0.072 +0.092 -0.133 -0.114 -0.209 -0.188 -0.147
UB3LYP/6-31G*//UHF/3-21G* +0.054 +0.088 +0.008 +0.052 -0.119 -0.132 -0.139 -0.167 -0.156
UB3LYP/6-31G*//AM1
UB3LYP/6-31G*//MNDO
UB3LYP/6-31G*//MNDO
UB3LYP/6-31G*//MNDO
UB3LYP/6-31G*//UHF/3-21G* +0.076 +0.111 +0.059 +0.056 -0.228 -0.257
UB3LYP/6-31G*//AM1
•
•
-
-
model of 7
model of 7
-0.052 +0.088 +0.024 +0.081 -0.115 -0.137 -0.165 -0.206 -0.152
+0.006 +0.006 -0.144 -0.130 +0.005 +0.006 +0.061 +0.073 -0.003
-0.043 -0.041 +0.009 +0.006 +0.009 -0.011 +0.010 +0.008 -0.456
+0.084 +0.134 +0.106 +0.069 -0.193 -0.184 -0.098 -0.131 -0.201
•-
+
7
7
7
8
8
/Li (A)
•
-
•
+
/Li (B)
-
+
/Li (C)
•
•
-
-
-0.333 -0.101
-0.311 -0.123
+0.100 +0.141 +0.081 +0.077 -0.234 -0.264
Ta ble 4. Ca lcu la ted a H of th e Ar om a tic P r oton s of th e Ra d ica l Ca tion s of 1, 7, a n d 8 a n d Mod el System s of 5-7 (for
m od el system s of 5/6, See F igu r e 10)
aH/mT (for the labeling of the positions, see the formulas)
method
meta
meta′ meta′′ meta′′′ ortho
ortho′ ortho′′ ortho′′′ para
para′
1•
+
+
+
UB3LYP/6-31G*//UHF/3-21G* +0.106 +0.073 +0.106 +0.073 -0.153 -0.171 -0.171 -0.153 -0.239
•
1
1
UB3LYP/6-31G*//AM1
UB3LYP/6-31G* (SP)
+0.130 +0.092 +0.130 +0.092 -0.182 -0.201 -0.201 -0.182 -0.254
+0.079 +0.085 +0.079 +0.085 -0.182 -0.264 -0.264 -0.182 -0.218
•
90°
•
•
+
+
model of [5/6]
model of [5/6]
UB3LYP/6-31G*//UHF/3-21G* +0.098 +0.058 +0.036 +0.027 -0.130 -0.137 -0.115 -0.163 -0.173
UB3LYP/6-31G*//AM1
+0.110 +0.067 +0.056 +0.047 -0.144 -0.154 -0.137 -0.193 -0.182
•+
7
UB3LYP/6-31G*//UHF/3-21G* +0.099 +0.075 +0.119 +0.079 -0.156 -0.176 -0.151 -0.169 -0.251
UB3LYP/6-31G*//UHF/3-21G* +0.101 +0.076 +0.119 +0.079 -0.157 -0.178 -0.142 -0.168 -0.258
UB3LYP/6-31G*//AM1
UB3LYP/6-31G*//UHF/3-21G* +0.114 +0.086 +0.066 +0.054 -0.190 -0.260
UB3LYP/6-31G*//AM1
•
•
+
+
model of 7
model of 7
-0.118 +0.097 +0.122 +0.075 -0.184 -0.217 -0.143 -0.167 -0.289
•
•
+
+
8
8
-0.333 -0.142
-0.265 -0.007
+0.111 +0.084 +0.077 -0.115 -0.197 -0.213
B3LYP method. The thus-calculated a
order of magnitude as the values obtained from the
minimized geometry (Figures 8 and 9). It is, however,
H
possess the same
parameters of the radical anions and cations derived from
tetraphenylethene.
Ion P a ir in g in th e Ra d ica l An ion s. It is well
•
-
apparent that the a
H
obtained from the optimized geom-
established that the π system of 1 and its derivatives
strongly interacts with its counterions.1
9,20,23,51
Therefore,
etry represent a preferable agreement with the experi-
mental values. In addition, we began optimizations with
the diphenylmethylene fragments being orthogonal for
both radical ions. In each case no minimum could be
localized for an almost orthogonal arrangement of the two
diphenylmethylene moieties.
it was of interest in how far the interaction of the
(crowno)phane moiety affects the electron distribution in
•
-
•-
4 -8 . No particular interactions between the alkyli-
dene chain in 4 is expected whereas 5 and 7 (and with
Hence, we conclude that the bond lengths and angles
indicated in Figure 10 represent relevant geometrical
(
51) Cserhegyi, A.; J agur, G. J .; Szwarc, M. J . Am. Chem. Soc. 1969,
91, 1892.

8
814 J . Org. Chem., Vol. 63, No. 24, 1998
Barbosa et al.
ever, the X-ray structure determination of the dianion
2
-
+
1
/2 Na reveals a twist angle between the two diphen-
+
ylmethylidene moieties of 56° with the Na counterions
above the central (formal) CdC bond.52 This arrange-
ment is in line with the ion-pair geometry stated for the
monoanions.
Radical anions in which a crown-ether moiety is
attached to a π system have been described. It was
shown that the crown ether may well serve as a ligand
to metal cations.5
3-57
In these systems, like 10, an oxygen
F igu r e 11. Ion-pair structures A, B, and C of 7•-/Li+.
some limitations 8) with their 12-crown-4 type chain
+
could serve as specific ligands for Li and 6, resembling
+
1
8-crown-6, could be specific for K . Unexpectedly, the
atom carrying a partial negative charge serves as an
additional ligand to the metal cation which then is also
embedded in the crown-ether moiety. Such an arrange-
ment is not possible in the phanes 4-8. The negative
charge is exclusively delocalized in the tetraphenylethene
π system, and the phenyl rings are oriented in such a
way that a simultaneous interaction of the crown ether,
the π system as the ligands, and the alkali metal atom
is not possible. We conclude that the Coulombic attrac-
tion between the negatively charged π system and the
metal cation is too strong to be matched by the interac-
tion between the crown ether and the cation.
ESR data of the phane radical anions indicate no
+
+
+
significant alternations with Li , Na , or K as the
counterion. All ESR spectra remained temperature
invariable in the range from 193 to 273 K. As stated
above, the differences in the a
can be traced back to the reduced symmetry of the phanes
-8 in respect to parent 1 and to electronic effects of the
H
and their multiplicities
4
alkyl, alkoxy, and acyl substituents rather than to
changes of the geometry or spin-distribution.
To get insight into the association properties of the
crownophanes we chose 7 as an representative example
and calculated three different structures of the ion pair
•-
•-
+
+
7
/Li (using MNDO, see Table 3). In the first case Li
The solution structures of the radical cations closely
match the geometry recently established by X-ray dif-
was placed symmetrically above the two phenyl rings
carrying the crown ether moiety (A), in the second case
above the two other 1,2-diphenyl groups (B), and in the
third case above the central double bond (C, Figure 11).
We then performed single-point calculations of the Fermi
contacts with the B3LYP functional. Whereas the posi-
tioning of the counterion at the benzene rings causes a
marked redistribution of spin and charge to the positions
5
8
fraction analysis of crystalline tetraanysilethene where
the ethenic bond is twisted by 30.5° and phenyl-ring
torsion amounts to 33° (cf. Figure 10). For the radical
cations no specific association with counterions could be
established by our experimental techniques; however, it
3+
is conceivable that metal cations such as Tl can be
embedded inside the crown ether moiety which is remote
from the positively charged π system.
+
next to the Li cation, a spin distribution which is
comparable to that found on the hyperfine time scale of
the ESR experiment emerges when the cation resides at
the central (formal) double bond (structure C).
Ackn owledgm en t. We thank the Volkswagen Foun-
dation for financial support and Prof. C. Bolm (Aachen,
Germany) for his encouragement. F.B. is indebted to
Prof. B. Giese (Basel) for his support.
Gen er a l Con clu sion s
According to our results, the radical anions and the
radical cations of parent 1 and phane derivatives 4-8
have a rather similar structure. The central ethenic bond
is twisted by ca. 35°, and the phenyl groups tend to adopt
a conformation in which they achieve a more coplanar
arrangement than in the neutral stage (enhancement of
electron delocalization). This is corroborated by the good
correlation between the calculated (UB3LYP/6-31G*//
J O980920O
(52) Bock, H.; Ruppert, K.; N a¨ ther, C.; Havlas, Z.; Herrmann, H.-
F.; Arad, C.; G o¨ bel, I.; J ohn, A.; Meuret, J .; Nick, S.; Rauschenbach,
A.; Seitz, W.; Vaupel, T.; Solouki, B. Angew. Chem., Int. Ed. Engl. 1992,
3
1, 550.
53) Hashimoto, K.; Togo, H.; Morihashi, K.; Yokoyama, Y.; Kikuchi,
(
O. Bull. Chem. Soc. J pn. 1991, 64, 3245.
(54) Tajima, K.; Miyoshi, M.; Furutani, M.; Fujimura, Y.; Mukai,
K.; Ishizu, K. Bull. Chem. Soc. J pn. 1989, 62, 2061.
AM1 or UB3LYP/6-31G*//UHF/3-21G*) a
H
and the shapes
of the frontier orbitals. Moreover, the experimental data
underline the validity of the pairing principle. The
rotation of the phenyl groups is restricted, mirrored by
the lack of the pairwise equivalence of the ortho and meta
protons. Peculiarly, it was reported that crystallization
of the tetraphenylethene radical anion had failed; how-
(55) Bock, H.; Hauck, T.; Naether, C.; Havlas, Z. Angew. Chem., Int.
Ed. Engl. 1997, 36, 638.
(56) Bock, H.; Hierholzer, B.; Voegtle, F.; Hollmann, G. Angew.
Chem. 1984, 96, 74.
(
(
57) Bock, H.; Herrmann, H. F. J . Am. Chem. Soc. 1989, 111, 7622.
58) Rathore, R.; Lindeman, S. V.; Kumar, A. S.; Kochi, J . K. J . Am.
Chem. Soc. 1998, 120, 6931.