An exceptional red shift of emission maxima upon fluorine substitution

Article abstract of DOI:10.1021/jo025592d

The effect of perfluorination on photophysical properties was investigated through synthesis and photophysical characterization of two isostructural donor-acceptor-donor dye molecules. The synthesis of two versatile fluorinated benzene compounds, 1,4-difl

Full text of DOI:10.1021/jo025592d

An Excep tion a l Red Sh ift of Em ission Ma xim a u p on F lu or in e
Su bstitu tion
Frederik C. Krebs* and Holger Spanggaard
The Danish Polymer Centre, RISØ National Laboratory, P.O. Box 49, DK-4000 Roskilde, Denmark
frederik.krebs@risoe.dk
Received February 5, 2002
The effect of perfluorination on photophysical properties was investigated through synthesis and
photophysical characterization of two isostructural donor-acceptor-donor dye molecules. The
synthesis of two versatile fluorinated benzene compounds, 1,4-difluoro-2,5-diperfluorooctylbenzene
(1) and 1,4-dibromo-2,5-difluoro-3,6-diperfluorooctylbenzene (2), is presented. The X-ray structure
of 2 has been determined and shows that the perfluorinated octyl chains segregate from the benzene
rings in the solid state, giving rise to a layered structure. The further synthesis through Suzuki
coupling reactions using 4-formylbenzeneboronic acid with (2) and 1,4-dibromo-2,5-dioctylbenzene
(3) gave, respectively, 1,4′′-diformyl-2′,5′-difluoro-3′,6′-diperfluorooctyl-p-terphenylene (4) and 1,4′′-
diformyl-2′,5′-dioctyl-p-terphenylene (5). The condensation of the dialdehydes 4 and 5 with 9,10-
phenanthrenequinone and ammoniumbicarbonate in glacial acetic acid gave the dye molecules 1,4′′-
bis(1H-phenanthro[9,10-d]imidazol-2-yl)-2′,5′-difluoro-3′,6′-diperfluorooctyl-p-terphenylene (6) and
1
,4′′-bis(1H-phenanthro[9,10-d]imidazol-2-yl)-2′,5′-dioctyl-p-terphenylene (7), respectively. The UV-
vis spectra of the two molecules are nearly identical, whereas the fluorescence spectra are very
different. Compound 7 shows blue fluorescence with little solvent dependence (λemission ) 410 nm in
THF, CH Cl , and hexane), whereas compound 6 shows a highly solvent-dependent emission
2 2
wavelength (λemission ) 583 nm in THF, λemission ) 560 nm in CH Cl , and λemission ) 450 nm in hexane).
2 2
The fluorescence red shift of compound 6 in a series of solvents with different polarity is discussed
using the Lippert-Mataga equation. Fluorescence lifetime and quantum yields were also
determined. Ultraviolet photoelectron spectroscopy (UPS) was performed on thin films of compound
6
and 7 on a gold substrate. The observed ionization potential was 4.55 eV in both cases.
In tr od u ction
the LUMO resides on the aromatic system to which the
H-phenanthro[9,10-d]imidazol-2-yl group is attached. In
1
The 1H-phenanthro[9,10-d]imidazol-2-yl group is worth
considering as a substituent in the field of molecular
materials. It has many desirable properties such as good
stability, ease of introduction into any molecular system
containing an aromatic aldehyde, use as a chromophore
with high extinction coefficient, readily tuneable absorp-
tion wavelength, and fluorophoric properties and is
desirable as a large planar synthetic building block in
supramolecular chemistry. Aside from the early reports
this paper we describe the synthesis and photophysical
properties of two new compounds of this class that are
isostructural terphenylenes with 1H-phenanthro[9,10-d]-
imidazol-2-yl groups. The central part of the terphenylene
group was prepared in two versions, substituted with
either hydrogen or fluorine, as shown in Scheme 1. The
object of this paper is to compare the differences in
properties obtained upon substitution of the hydrogen by
fluorine.
1
of its synthesis straight from phenanthrenequinone and
an aromatic aldehyde, the system has received little
interest. Recently, the use of the 1H-phenanthro[9,10-d]-
imidazol-2-yl group because of its supramolecular proper-
ties has been reported in a molecular tweezer system for
Resu lts a n d Discu ssion
Syn th esis. The perfluoroalkylation of 1,4-dibromo-2,5-
difluorobenzene using a procedure modified from that
given in ref 4 with perfluorooctyliodide and copper dust
in dry dimethylsulfoxide proceeds in very good yield to
2
explosives and as a fluorophore in a superradiant laser
3
dye as a result of its photophysical properties. In the
latter case it has been found that the HOMO resides on
the 1H-phenanthro[9,10-d]imidazol-2-yl group and that
give 1. In analogy with the bromination of pentafluo-
5
robenzene, bromination of 1 in 65% oleum using AlBr
3
as a catalyst gives initially a heterogeneous mixture that
after reflux for 6 h becomes homogeneous. When con-
(
1) (a) Krieg, B.; Manecke, G. Z. Naturforsch., B: Anorg. Chem. Org.
6
Chem. Biochem. Biophys. Biol. 1967, 22, 132. (b) Neunhoeffer, O.;
Krieg, B. Z. Naturforsch., B: Anorg. Chem. Org. Chem. Biochem.
Biophys. Biol. 1966, 21, 536. (c) Cook, A. H.; J ones, D. G. J . Chem.
Soc. 1941, 278. (d) Steck, E. A.; Day, A. R. J . Am. Chem. Soc. 1943,
(4) McLoughlin, V. C. R.; Thrower, J . Tetrahedron 1969, 25, 5921-
5940.
6
5, 452.
(
2) Krebs, F. C.; J ørgensen, M. J . Org. Chem. 2001, 66, 6169-6173.
(5) Knunyants, I. L.; Yakobsen, G. G. Synthesis of Fluoroorganic
Compounds; Springer-Verlag: Berlin, Heidelberg, New York, Tokyo,
1985.
(3) Krebs, F. C.; Lindvold, L. R.; J ørgensen, M. Tett. Lett. 2001, 42,
6
1
753-6757.
0.1021/jo025592d CCC: $22.00 © 2002 American Chemical Society
Published on Web 09/14/2002
J . Org. Chem. 2002, 67, 7185-7192
7185

Krebs and Spanggaard
SCHEME 1
SCHEME 2
F IGURE 1. An illustration of the layered crystal structure
observed for compound 2 (bromine atoms are red and fluorine
atoms are green).
tinuing the reaction for another 14 h the dibrominated
product 2 is obtained, as shown in Scheme 2.
carbon atoms. This prompted us to try Suzuki-type
reactions on compound 2, and this was found to be
The products are poorly soluble in common solvents
such as methylenechloride, chloroform, and ether and
moderately soluble in boiling toluene and boiling tet-
rahydrofurane. Characterization of the products using
possible under standard conditions using (PPh
3 2 2
) PdCl
as catalyst and Na CO (aq) as the base in toluene at
2
3
standard NMR techniques is essentially limited to 19
reflux temperature, as shown in Scheme 3. Further
reaction of the dialdehydes 4 and 5 with ammoniumb-
icarbonate and phenanthrenequinone in glacial acetic
acid gave 6 and 7.
F
NMR. Compound 1 contains only two protons and 2
1
contains none, which means that H NMR gives little
information and traditional 13C NMR without special
broadband C-F decoupling schemes gives rise to a
myriad of signals due to the magnitude of the various
P h otop h ysica l Stu d ies. The dye molecules 6 and 7
have very similar UV-vis spectra in both apolar (toluene)
and polar (THF) solvents, as shown in Figure 2. The main
differences between 6 and 7 in the absorption spectra are
a slightly higher extinction coefficient, a small blue shift
of the first maximum by 2 nm, and a small shoulder at
the absorption edge for the fluorine compound (compound
carbon-fluorine coupling constants. The recording of 13
C
NMR spectra of the neat compounds in the molten state
with a capillary containing DMSO-d ) for 1-2 days gave
only limited information. Elemental analysis was possible
with addition of V to avoid combustion problems, and
(
6
2
O
5
standard mass spectroscopy was also applied. A crystal
structure was obtained that not only confirmed the
molecular structure of 2 but also shed some light on the
solid state behavior of the perfluorinated side chains, as
few reports are available for this type of compound. The
fluorinated side chains segregate, and a layered structure
is observed, as shown in Figure 1.
6
). We believe that the smaller extinction coefficient
arises because of a steric effect due to the fluorine atoms
at the central benzene ring in the terphenylene system.
Although neither 6 nor 7 is believed to acquire a planar
conformation in the terphenylene system, the steric effect
of the fluorine atoms in 6 makes the torsion angle
between the central benzene ring and its aryl substitu-
ents larger than it is in 7, as the geometry optimization
shows (59° for 6 and 28° for 7). Further evidence of a
nonplanar geometry is found in the comparison of 6 and
7 with 4,4′-bis(1H-phenanthro[9,10-d]imidazol-2-yl)bi-
phenyl, which is known to acquire a planar geometry
Two comparable crystal structures could be found in
the literature, one for a perfluorohexyl substituted aro-
matic compound and one for a perfluoroheptyl substi-
tuted aromatic system. Both examples show the same
tendency for the perfluorinated side chains to segregate
from the rest of the molecular system in the crystal.7
All attempts to perform ortholithiation or metal-
halogen exchange on this series of molecules using
3
leading to absorption maxima beyond 400 nm.
Although both spectra are similar in both THF and
toluene, it is noteworthy that the extinction coefficient
for 7 is much more dependent on the polarity of the
solvent. The extinction coefficient for both 6 and 7
decreases on going from polar THF to the less polar
toluene, but while 6 only shows a slight decrease by about
10% the extinction coefficient of compound 7 decreases
by a much more significant 40%, as can be seen from the
data presented in Table 1. This is consistent with
observations made for a system with a similar donor-
acceptor-donor geometry.11 The shoulder that is observed
in 6 is slightly more pronounced in apolar solvents (vide
supra). To substantiate the observations in the UV-vis
spectra the isolated 1H-phenanthro[9,10-d]imidazol (8)
was prepared and further comparison was made with
2-phenyl-1H-phenanthro[9,10-d]imidazol (9) and 2-(4-
n
t
n
standard bases ( BuLi, BuLi, BuLi/TMEDA, PhLi) in
common solvents (ether, THF, dioxane, ether/hexane)
failed in our hands and we ascribe these problems to a
lack of suitable solvents for these molecules at low
temperature. The most successful attempt was obtained
n
in ether/hexane (1:l) at 0 °C using BuLi, where formyl-
ated products could be obtained after quenching with
dimethylformamide. A significant degree of n-butylation
was, however, also observed and believed to take place
on both the aromatic carbon atoms and on the side chain
(6) Attempts to isolate the product at this point gave mainly the
monobrominated species; in our hands it was not possible obtain a pure
sample free from 1 and 2.
(7) Kromm, P.; Allouchi, H.; Bideau, J .-P.; Cotrait, M. Mol. Cryst.
Liq. Cryst. Sci. Technol., Sect. A 1994, 257, 9. Kromm, P.; Bideau, J .-
P.; Cotrait, M.; Destrade, C.; Nguyen, H. Acta Crystallogr. 1994, C50,
1
12.
7
186 J . Org. Chem., Vol. 67, No. 21, 2002

Emission Red Shift upon Fluorine Substitution
SCHEME 3
F IGURE 2. The UV-vis spectra of compounds 6 and 7 in
THF and toluene.
F IGURE 3. The UV-vis spectrum of the reference compound
8 in solvents of different polarity.
TABLE 1. λm a x (n m ) Va lu es a n d Associa ted Extin ction
SCHEME 4
-
1
-1
Coefficien ts E (M cm , in br a ck ets) fr om UV-vis
Sp ectr a of 6 a n d 7 in THF a n d Tolu en e
6
(toluene)
6 (THF)
7 (toluene)
7 (THF)
3
3
3
68 (48500)
50 (41500)
37 (42500)
368 (54500)
350 (47750)
335 (52500)
368 (33750)
350 (35250)
338 (36750)
370 (52250)
350 (54250)
338 (61750)
biphenyl)-1H-phenanthro[9,10-d]imidazol (10). Although
the synthesis of 1H-phenanthro[9,10-d]imidazol has been
reported, no analytical data can be found in the literature
aside from the melting point and elemental analysis. The
condensation reaction between phenanthrenequinone
and an aromatic aldehyde in acetic acid containing a
source of ammonium proceeds quite well, and this led
us to attempt the condensation reaction between phenan-
threnequinone and hexamethylenetetramine or aliphatic
aldehydes that has been described. We however failed
to produce the right product. Finally, the photochemical
reaction of diphenylimidazole and iodine in benzene was
employed, giving the isolated phenanthroimidazole moi-
ety 8 as shown in Scheme 4. It is noteworthy that while
many references including recent ones can be found, the
NMR data have not been presented.1 Therefore the data
were included in the supplementary section.
,8
The UV-vis spectrum of 8 (doubly sublimed) shows
features similar to that reported for phenanthro[9,10-d]-
oxazole9 except that there are weak transitions with
-1
-1
extinction coefficients of the order of 1000-2000 M cm
in the 320-350 nm range as shown in Figure 3 and Table
. The observation of these transitions was substantiated
(8) (a) Isikdag, I.; Ucucu, U.; Ozdemir, A.; Meric, A.; Ozturk, Y.;
Aydin, S.; Ergun, B. Boll. Chim. Farm. 1999, 138, 453-456. (b)
Purushothaman, E.; Rajasekharan-Pillai, V. N. Indian J . Chem. B.
2
1
989, 28, 290-293. (c) Oda, K.; Sakai, M.; Tsujita, H.; Machida M.
by ZINDO calculations (vide supra).
Synth. Commun. 1997, 27, 1183-1189.
(
9) Padwa, A.; Ku, H.; Mazzu, A. J . Org. Chem. 1978, 43, 381-387.
The weak transitions in the 320-350 nm region were
originally considered to be n-π* transitions, both because
of their low intensity and because it has been reported
that these transitions were n-π* and could be suppressed
(
10) Baumann, W.; Bischof, H.; Fr o¨ hling, J . C.; Brittinger, C, Rettig,
W.; Rotkiewicz, K. J . Photochem. Photobiol., A 1992, 64, 49.
11) Strehmel, B.; Sarker, A. M.; Malpert, J . H.; Strehmel, V.;
Seifert, H.; Neckers, D. C. J . Am. Chem. Soc. 1999, 121, 1226-1236.
(
J . Org. Chem, Vol. 67, No. 21, 2002 7187

Krebs and Spanggaard
TABLE 2. λm a x (n m ) Va lu es a n d Associa ted Extin ction
-
1
-1
Coefficien ts E (M cm , in br a ck ets) fr om UV-vis
Sp ectr a of Com p ou n d 8 in EtOH, THF , a n d Tolu en e
EtOH
THF
toluene
3
3
3
3
2
50 (1900)
35 (1600)
20 (950)
00 (7850)
80 (14800)
355 (2050)
338 (1750)
325 (1050)
303 (8500)
282 (16200)
354 (1400)
337 (1150)
322 (750)
304 (5850)
283 (9450)
TABLE 3. λm a x (n m ) Va lu es a n d Associa ted Extin ction
-
1
-1
Coefficien ts E (M cm , in br a ck ets) fr om UV-vis
Sp ectr a of 8, 9, a n d 10 in THF
8
9a
10a
3
3
3
3
55 (2050)
38 (1750)
25 (1050)
03 (8500)
364 (10600)
347 (12200)
327 (22100)
314 (21400)
372 (29850)
352 (29950)
340 (42650)
b
F IGURE 5. The emission spectra obtained for 8, 9, and 10 in
THF under magic angle conditions. All spectra are normalized
for easy comparison.
a
Values taken from ref 3 and rounded to nearest 50. b It is
noticeable that although no distinct peaks are observed there is a
shoulder indicating the presence of at least one peak.
F IGURE 6. The emission spectra obtained for 6 and 7 in polar
(THF) and nonpolar (hexane) solvents under magic angle
conditions. All spectra are normalized for easy comparison.
Notice the extreme polarity dependence of the emission
maximum in the case of 6.
F IGURE 4. The absorption spectra of the reference com-
pounds 8, 9, and 10 in THF.
9
. Compound 9 exhibits a slight bathochromic shift in
the emission as expected. Compound 10 however does
not show fine structure, which could be indicative of
solvent rearrangements and changes in the molecular
geometry upon excitation. The emission spectrum of 10
is similar to the emission spectrum of 7 (Figure 6), and
we interpret this as an indication that in both compounds
the excitation is localized at the phenanthroimidazol
moiety. Otherwise we would have expected an increased
bathochromic shift in the absorption spectrum compared
to that of 10. This then supports the concept that upon
excitation the phenanthroimidazol moiety can be subject
to either a planarization or partial planarization with the
phenylene ring upon which it is attached and with the
central phenylene group. In the case of a local excitation
the emission spectrum shows fine structure as found for
8 and 9. In a more extended system like 10 a planariza-
tion event (or solvent rearrangement) during the transi-
tion takes place. The transition can thus still be consid-
ered as one of local character and the emission is now
in hydrogen donor solvents such as ethanol.8b As seen in
Figure 3 and Table 2 the absorption spectrum in ethanol
shows only minor changes in the magnitude of the
extinction coefficient and a hypsochromic shift of 5 nm.
The introduction of a phenyl or biphenyl substituent in
the 2-position of the 1H-phenanthro[9,10-d]imidazol
moiety as for compound 93 and 103 leads to a strong
increase in the magnitude of the extinction coefficient,
indicating π-π* character.
The UV-vis spectra show essentially the same fea-
tures subjected to a bathochromic shift as summarized
in Table 3. The hyperchromic shifts associated with all
of the peaks are also given in Table 3, and the UV-vis
spectra are shown in Figure 4. We therefore assume that
the transitions in the 320-350 nm region are π-π*
transitions.
In the emission spectra of the reference compounds
(shown in Figure 5) fine structure is observed for 8 and
7
188 J . Org. Chem., Vol. 67, No. 21, 2002

Emission Red Shift upon Fluorine Substitution
lacking fine structure. Only two of the phenylene units
in the terphenylene system are thus believed to be
involved in the excitation and emission processes.
The emission spectrum of 6 and 7 in THF and hexane
is shown in Figure 6. In all cases the spectra were broad
and there was no sign of fine structure. The emission
maximum of 6 showed a large red shift going from
nonpolar to polar solvents. This is in marked contrast to
the emission spectrum of 7, which shows virtually no
solvent dependence.
We ascribe this behavior to two different mechanisms.
In the case of 7 we have a local excitation involving some
rotation or planarization of the central phenylene moiety,
which gives rise to a lack of fine structure in the emission
spectrum. The solvent polarity dependence of the quan-
tum yield for the emission in 7 only shows a slight
decrease. For 6 the initial excitation can either be of the
same local character as for 7 or it could be a direct charge-
transfer transition connected to the shoulder in 6 that
lies underneath the dominant π-π* transitions. How-
ever, the strongly electron-withdrawing perfluorinated
central phenylene unit probably makes a different path-
way available where some degree of charge transfer is
possible, and the position of the emission maximum now
shows a strong dependence on solvent polarity. Further-
more, the quantum yield for the emission process de-
creases severely when solvent polarity is increased, as
expected.11 Typically there is a difference between the
ground state dipole and the excited state dipole of a
molecule. Upon excitation the solvent molecules will
rearrange according to the newly formed dipole vector
and thereby stabilize the excited state. In general the
more polar the solvent the more the excited state will be
stabilized. The effect is an increased red shift of the
F IGURE 7. The Lippert-Mataga plot for compound 6 in
dichloromethane/hexane mixtures (triangles with linear fit)
and for ethanol/hexane mixtures (squares) showing a nonlinear
relationship. Additionally the single points are obtained from
solutions in pure solvents with different dielectric constants.
TABLE 4. Qu a n tu m Yield s a n d Lifetim es a t Em ission
Ma xim u m (n s, in br a ck ets) for Com p ou n d s 6 a n d 7 in
Ar gon -Dega ssed Solven ts
compound
CH2Cl2
ethanol
6
7
0.36 (4.4)
0.68 (1.2)
0.024 (1.10)
0.45 (0.85)
will cancel each other to some extent. This means that
the observed red shift is less than should be expected if
the dipoles were isolated. Strehmel et al. have also
applied the Lippert-Mataga equation to molecules with-
out a net dipole moment but, as in our case, with a
quadrupole.11 For 6 in ethanol/hexane mixtures ν
is a
emission spectrum in more polar solvents, which is
f
described by the Lippert-Mataga equation,10
nonlinear function of ∆f, suggesting a specific interaction
between ethanol and 6, possibly H-bonding between
ethanol and the amine or imine functionality on the 1H-
phenanthro[9,10-d]imidazol-2-yl group. The effect of
ethanol on 6 can also be seen in the quantum yields and
lifetimes presented in Table 4.
1
2
FC
g
V ) C -
µ (µ - µ )∆f
(1)
F
1
e
e
3
hc4πꢀ0
a
FC
where C
1
is a constant, µ and µg are the dipole
e
moments in the excited and ground state, respectively,
and FC denotes that the ground state has an unrelaxed
Franck-Condon geometry. ∆f is the solvent polarity and
is given by eq 2 where ꢀ is the dielectric constant and n
is the optical refractive index.
In dichloromethane/hexane mixtures ν is a linear
f
-
1
function of ∆f with a slope of -21300 cm . Individual
points obtained for solutions in pure solvents were also
obtained, and a linear fit to these data gave a slope of
-
1
-19000 cm
.
The slope is relatively steep, indicating that the
quadrupole moment in the excited state is substantially
larger than the ground state quadrupole. Further, the
slope is substantially steeper than that obtained for
donor-π-acceptor molecules such as dialkylaminonaph-
2
ꢀ
- 1
(n - 1)
∆
f )
-
(2)
2
2
ꢀ + 1
(
2n + 1)
Equation 1 assumes the presence of a molecular dipole.
However 6 does not have a dipole moment, since it
contains a center of symmetry, but a quadrupole moment
with a charge of -2δ on the central terphenylene unit
and of +δ on each of the two 1H-phenanthro[9,10-d]-
imidazol-2-yl groups. Despite the lack of a dipole moment
the red shift of 6 in dichloromethane/hexane mixtures
can be described by the Lippert-Mataga equation (see
Figure 7). We speculate that because 6 is a rather large
molecule its quadrupole moment can be approximated
by two isolated dipole moments. This approximation is
most valid around the 1H-phenanthro[9,10-d]imidazol-
-
1
thalenesulfonamides (-10600 cm ) and coumarins
-
1 12
(-11300 cm
)
but similar to the slope obtained for a
series of donor-π-acceptor molecules forming intramo-
lecular charge transfer states upon excitation, such as
dimethylaminobenzonitriles (from -15000 to -24200
-
1
-1
8
cm ) and dimethylaminobenzoic esters (-27200 cm ).
This suggests that there is a substantial charge transfer
upon excitation. The ground state of 6 is expected to have
a strong quadrupole due to the powerful electron-
withdrawing effect of the fluorine atoms on the central
2
-yl unit and the least valid around the center of the
(
12) J ager, W. F.; Volkers, A. A.; Neckers, D. C. Macromolecules
molecule (the terphenylene unit) where the two dipoles
1995, 28, 8153.
J . Org. Chem, Vol. 67, No. 21, 2002 7189

Krebs and Spanggaard
terphenylene moiety. Semiempirical calculations on ge-
ometry optimized molecular structures (using MM2 with
a CVFF force field ) of 6 and 7 using the ZINDO
ferred to as twisted intramolecular charge transfer or
TICT.15 We speculate that such a TICT mechanism may
be responsible for the spectral broadening. The photo-
electron spectroscopy data show that the ionization
potentials are identical for 6 and 7. The only difference
is a slight shift of the energy levels to lower energies for
compound 6. Under the assumption that the samples are
free from impurities and the crystallites are randomly
oriented at the surface of the gold substrate, the Fermi
levels in 6 and 7 align with the Fermi level of the gold
substrate, making this observation possible (shown in
figure S1 in the Supporting Information). It is also
evident that the magnitude of the ionization potential is
identical for 6 and 7, but the relative position of the
energy levels is indeed lower for 6 as a result of the
inductive effect of the fluorine atoms. These results
confirm the similarity of the experimental electronic
structure in the solid state.
1
3
13
program (developed by M. Zerner with INDO/1 param-
etrization using CIS) showed that the HOMO is located
on the 1H-phenanthro[9,10-d]imidazol-2-yl group while
the LUMO is located on the central benzene core in the
terphenylene unit. Thus the effect of excitation is to
transfer charge to the central part of the molecule,
thereby increasing the existing quadrupole. The processes
can be summarized in the following scheme shown below
(taken from ref 11):
absorption
fluorescence
LE yfluorescencez LE* f CT* yabsorption z CT
In the case of compound 6 the emission comes from the
excited charge transfer state (CT*) independent of the
state it was in upon excitation. In compound 7, however,
there is no process that leads to the CT* state and
emission comes solely from the locally excited state (LE*).
The calculations also show that the HOMO and the
LUMO of 7 are very similar to those of 6. In the case of
Con clu sion s
We have presented the synthesis of two isostructural
donor-acceptor-donor dye molecules, one fully hydro-
genated and one fluorinated on the acceptor part of the
molecule. We show by photochemical means that while
the electronic structure of the two compounds are very
similar for the two compounds, as established by UV-
vis and photoelectron spectroscopy, the emission proper-
ties are very different. The effects were very noticeable
for the fluorinated molecule where the positions of
hydrogen-to-fluorine substitution were chosen to stabilize
the excited state. As expected this resulted in a large
dependence of the emission maximum on the solvent
polarity, whereas the compound containing only hydro-
gen showed no significant dependence on solvent polarity.
7
however the charge transfer to the central terphenylene
unit upon excitation is not stabilized by the inductive
effect of the fluorine atoms and the quadrupole is
expected to be much smaller for 7, thus favoring the local
excitation. By the same argument the ground-state
quadrupole of 7 is expected to be smaller than that of 6.
Equation 1 shows that the emission frequency varies with
the square of the excited-state dipole moment and has a
linear dependence with respect to the ground-state dipole
moment; 6 has the strongest excited state quadrupole and
a much larger difference in charge distribution between
the ground and excited state. Consequently 6 shows a
much stronger red shift in the emission wavelength than
Exp er im en ta l Section
7
(
. When the emission spectra of 6 in hexane and CH
and THF) are compared, it is obvious that the latter is
very broad, but it also seems that the shoulder in hexane
around 500 nm) is missing in CH Cl (and THF). This
may indicate that emission takes place from one state
in hexane and another in CH Cl (and THF).
2
Cl
2
All reagents were commercially available unless otherwise
stated. 1,4-Dibromo-2,5-dioctylbenzene was prepared as de-
scribed in ref 16, and 4-formylphenylboronic acid was prepared
as described in ref. 17. The solubility of compound 6 in DMSO-
(
2
2
1
3
6
d at 440 K was too low to obtain a C NMR spectrum. The
1
19
H and F NMR spectra were obtained by long acquisition
periods. High resolution (HRMS) mass spectrometry was
performed using using a wavelength of 337 nm for matrix-
2
2
To examine the possible presence of an excited state
reaction, time resolved emission spectroscopy (TRES) was
performed on 6 in CH
progressive red shift of the emission with time as
expected. This is a consequence of increasing solvent
reorientation around 6. More importantly the spectra
showed a slight broadening (FWHM) by about 400 cm
during the first 1-2 ns after excitation. This might
indicate that a second state responsible for the broaden-
ing is appearing fairly quickly. In the ground state most
conjugated π-systems will try to maintain a planar
structure, but in our case the conjugated system is forced
out of the ideal planar arrangement as a result of the
steric repulsions from the bulky perfluorooctyl substit-
uents. Upon excitation the high degree of charge transfer
effectively disrupts the conjugated system whereby the
central terphenylene unit is free to rotate to minimize
the steric repulsionssthis mechanism is normally re-
assisted laser desorption ionization (MALDI). The peak at m/z
+
2
Cl
2
. The investigations showed a
273.0393 of [M - H
2
O + H] ions of the matrix compound (2,5-
dihydroxybenzoic acid, DHB) was used as an internal calibra-
tion standard.
2
,5-Diflu or o-1,4-b is(p er flu or ooct yl)b en zen e (1). Cu-
bronze (50 g) was stirred in dry DMSO (400 mL, dried by
distillation from CaH ) and degassed with argon. Perfluorooc-
-
1
2
tyl iodide (200 g, 0.366 mol) was added, and the mixture was
heated on an oil bath (140 °C bath temperature). After 45 min
2,5-difluoro-1,4-dibromobenzene (47.5 g, 0.174 mol) was added,
and the mixture was left overnight in the oil bath. The mixture
was cooled to room temp for 2 h. A lower layer separated,
which became solid. Aqueous ammonia (0.2 L) was added. and
the supernatant was decanted. The solid lump was washed
with several portions of aqueous ammonia. The lump was
dissolved in boiling toluene (3 L) and filtered, giving a light
yellow filtrate with a white precipitate. The mixture was
cooled. The colorless product was filtered, washed with toluene
1
4
(
15) Rettig, W. Top. Cur. Chem. 1994, 69, 253.
(13) Insight II User Guide, October 1995. Biosym/MSI: San Diego,
(16) Rehahn, M.; Schlueter, A.-D.; Feast, J . W. Synthesis 1988, 5,
1
995.
386-388.
(14) R o¨ cker, C.; Heilemann, A.; Fromherz, P. J . Phys. Chem. 1996,
(17) Park, P. C.; Yoshino, K.; Tomiyasu, H. Synthesis 1999, 12,
2041-2044.
1
00, 12172.
7
190 J . Org. Chem., Vol. 67, No. 21, 2002

Emission Red Shift upon Fluorine Substitution
(
2 × 500 mL) and methylene chloride (2 × 500 mL), and dried.
TABLE 5. Cr ysta llogr a p h ic Da ta for Com p ou n d 2
This gave 1 as a colorless material in 76% yield (125 g): mp
1
formula
formula wt
crystal system
space group
Z
C22F36Br2
1108.04
monoclinic
P21/c
1
11-112 °C; H NMR (250 MHz, CDCl , 330 K, TMS) δ 7.5
19
3
3
4
(
dd, J (H,F) ) 9 Hz, J (H,F) ) 7 Hz, 2H); F NMR (235 MHz,
CDCl , 330 K, C ) δ -122.2 (m, 4F), -118.7 (m, 4F), -117.7-
m, 16F), -110.6 (m, 2F), -105.9 (m, 4F), -77.2 (m, 6F). Anal.
Calcd for C22 36: C, 27.81; H, 0.21. Found: C, 27.91; H, 0.13.
,4-Dibr om o-2,5-diflu or o-3,6-per flu or ooctylben zen e (2).
Aluminum foil (1 g) was cut into small pieces and placed in a
L round-bottomed flask with a large stirring bar. The flask
3
6 6
F
(
2
a, Å
22.089(8)
5.863(2)
11.966(5)
90
99.159(9)
90
2
H F
b, Å
c, Å
1
R, deg
1
â, deg
γ, deg
was flushed with argon. Bromine (10 mL) was added in one
portion, and the mixture was stirred vigorously under Ar until
the aluminum had reacted (CAUTION!). The aluminum foil
burns vigorously in the bromine vapor with a bright luminous
flame, and care has to be taken that the aluminum pieces do
not stick to the side of the flask but are kept stirring round.
When the reaction has subsided oleum (60%, 300 mL) was
added followed by bromine (70 mL, excess) and 1 (48 g, 0.05
mol). A condenser was fitted, and the mixture was stirred
under Ar in an oil bath with a temperature of 40 °C. Initially
the starting material floats inside the flask, but after 6 h the
mixture becomes homogeneous. After 20 h the mixture was
poured into ice (4 kg) and filtered. The product was washed
V, Å3
1529.9(10)
2.405
3
F, g cm
crystal dimensions, mm
0.75 × 0.75 × 0.006
type of radiation
Mo KR
µ, cm-1
2.890
T, K
120(2)
no. of reflections
unique reflections (I > 2σ)
18902
1706
2
R(F), Rw(F ) all data
0.0992, 0.2788
0
.57 mmol) and 9,10-phenanthrenequinone (1 g. 4.8 mmol)
were mixed in AcOH (50 mL), and NH HCO (2 g, excess) was
4
3
added. The mixture was heated to reflux, giving a light orange
and clear solution. After 15 min the mixture becomes cloudy,
and a light yellow precipitate forms. After 3 h the reaction
was stopped, and the mixture was cooled overnight. The solid
was filtered, washed with light petroleum, and recrystallized
with water (2 × 500 mL), Na
2
S
2
O
6
(aq) (1 M, 2 × 500 mL) and
recrystallized from toluene (1.5 L). This gave compound 2 in
1
9
8
3
1
2% yield (46 g): mp 131-132 °C; F NMR (235 MHz, CDCl
30 K, C ) δ -122.2 (m, 4F), -118.7 (m, 4F), -117.7 (m,
2F), -116.3 (m, 4F), -100.4 (m, 4F), -89.6 (m, 2F), -77.2
36: C, 23.85; H. 0.00. Found:
3
,
6
F
6
(m, 6F). Anal. Calcd for C22Br
2
F
from toluene containing a little Et
3
N. This gave 6 in 45% yield
-
l
C, 23.83; H, 0.00.
(0.4 g). DSC data for 6: peak at 349.27 °C, ∆H ) 76437 J mol ;
1
4
,4′′-Difor m yl-2′,5′-diflu or o-3′,6′-per flu or ooctylter ph en -
ylen e (4). Compound 2 (5 g, 4.5 mmol), 4-formylphenylboronic
acid (5 g, excess), Na CO (14 g), water (150 mL), and toluene
300 mL) were mixed and degassed with argon. (PPh PdCl
300 mg, catalyst) was added, and the mixture was refluxed
6
H NMR (250 MHz, DMSO-d , 440 K, TMS) δ 7.6-7.8 (m,
3
3
12H), 8.4 (d, J (H,H) ) 8 Hz, 4H), 8.6 (d, J (H,H) ) 8 Hz, 4H),
8.8 (d, J (H,H) ) 8 Hz, 4H), 13.1 (bs, 2H); F NMR (235 MHz,
DMSO-d , 440 K, C (sealed capillary)) δ -121.2 (m, 4F),
3
19
2
3
(
(
3
)
2
2
6
6 6
F
-117.7 (m, 4F), -116.1 (m, 12F), 113.9 (m, 4F), 103.7 (m, 2F),
-96.1 (m, 4F), -76.1 (m, 6F); HRMS (MALDI, DHB): m/z
for 24 h. An additional amount of 4-formylphenylboronic acid
3 g, excess) was added with (PPh PdCl (300 mg, catalyst),
(
3
)
2
2
calcd for [C H26F36N + H]
+
1535.1655, found 1535.1644. Anal.
6
4
4
and reflux continued for another 24 h. The mixture was cooled
and left to crystallize. Filtering gave the product containing
darkly colored catalyst decomposition products. The solids
were boiled in toluene (400 mL) and filtered to give a near
colorless solution that was left to crystallize, giving pure 4 as
fine thin needle shaped crystals in 61% yield (3.2 g): mp 189-
Calcd for C64
H
26
F
36
N
4
: C, 50.08; H, 1.71; N, 3.65. Found: C,
5
0.00; H, 1.64; N, 3.41.
,4-Dioctyl-2,5-bis(4-(1H-p h en a n th r o[9,10-d ]im id a zol-
-yl)p h en yl)ben zen e (7). Compound 5 (0.4 g, 0.78 mmol),
,10-phenanthrenequinone (0.6 g, 2.9 mmol), and NH HCO
1
2
9
4
3
1
3
(2 g, excess) were mixed in AcOH (50 mL) and heated to reflux
for 3 h. After cooling the colorless product was filtered and
washed with acetone. Recrystallization from toluene contain-
3
1
90 °C; H NMR (250 MHz, CDCl
3
, 330 K, TMS) δ 7.4 (d, J
3
(
H,H) ) 8 Hz, 4H), 8.0 (d, J (H,H) ) 8 Hz,, 4H), 10.1 (s, 2H);
C NMR (63 MHz, CDCl
37.7, 191.8; F NMR (235 MHz, CDCl , 330 K, C ) δ -122.2
1
3
3
, 330 K, TMS) δ 129.9, 130.7, 137.0,
1
9
ing a little Et N gave 7 in 61% yield (0.42 g). DSC data for 7:
1
3
6 6
F
-
1 1
(
m, 4F), - 118.8 (m, 4F), -117.8 (m, 12F), -115.5 (m, 4F),
102.7 (m, 2F), -98.0 (m, 4F), -77.2 (m, 6F). Anal. Calcd for
: C, 37.33; H, 0.87. Found: C, 37.13; H, 0.76.
,4′′-Difor m yl-2,5-d ioctylter p h en ylen e (5). 1,4-Dibromo-
,5-dioctylbenzene (10 g, 21.7 mmol), 4-formylphenylboronic
acid (10 g, excess), and Na CO (10.6 g, 0.1 mol) were mixed
in toluene (150 mL) and water (100 mL), and the mixture was
degassed with argon. (PPh PdCl (350 mg, catalyst) was
peak at 331.21 °C, ∆H ) 94037 J mol ; H NMR (250 MHz,
DMSO-d , 300 K, TMS) δ 0.7 (t, 6H), 1.1 (m, 20H), 1.5 (m,
4H), 2.7 (t, 4H), 7.3 (s, 2H), 7.6-7.8 (m, 12H), 8.4 (d, J (H,H)
-
6
3
36 10 36 2
C H F O
3
3
)
8 Hz, 4H), 8.6 (t, J (H,H) ) 8 Hz, 4H), 8.9 (d, J (H,H) ) 8
13
4
2
6
Hz, 4H), 13.5 (s, 2H); C NMR (63 MHz, DMSO-d , 300 K,
TMS) δ 13.8, 22.0, 28.4, 28.5, 28.8, 30.7, 31.2, 32.0, 121.8,
121.9, 122.4, 123.7, 124.1, 125.1, 125.2, 125.3, 125.9, 126.9,
127.0, 127.1, 127.5, 127.7, 128.9, 129.5, 130.7, 137.0, 137.2,
139.9, 142.0, 148.9; HRMS (MALDI, DHB): m/z calcd for
2
3
3
)
2
2
added, and the mixture was refluxed for 24 h. An additional
portion of 4-formylphenylboronic acid (5 g, excess) and Pd-
(
+
[C H62N4 + H] 887.5053, found 887.5170. Anal. Calcd for
64
PPh
3
)
4
(350 mg, catalyst) were added, and the mixture was
64 62 4
C H N : C, 86.64; H, 7.04; N, 6.32. Found: C, 86.33; H, 6.87;
refluxed for 24 h. The reaction was stopped, and the organic
phase was separated and concentrated. The crude oil was
chromatographed on silica using toluene as eluent. Evapora-
N, 6.22.
Cr ysta llogr a p h y. General crystallographic data can be
found in Table 5. The crystal of 2 was drawn directly from
the mother liquour, coated with a thin layer of protecting oil,
and mounted on a glass fiber using grease and transferred
quickly to the cold stream of nitrogen on the diffractometer.
tion gave an oil that crystallized from petroleum, giving 5 in
1
2
3
4
8
5% yield (2.83 g): mp 61-63 °C; H NMR (250 MHz, CDCl
3
,
00 K, TMS) δ 0.8 (t, 6H), 1.2 (min, 20H), 1.5 (m, 4H), 2.6 (t,
3
3
H), 7.1 (s, 2H), 7.5 (d, J (H,H) ) 8 Hz, 4H), 7.9 (d, J (H,H) )
1
3
Hz, 4H), 10.1 (s, 2H); C NMR (63 MHz, CDCl
3
, 300 K, TMS)
The crystals were very thin plates and this impaired the
quality of the data. An almost complete sphere of reciprocal
space was covered by a combination of several sets of exposure
frames; each set with a different æ angle for the crystal and
each frame covering a scan of 0.3° in ω. Data collection,
integration of frame data, and conversion to intensities cor-
rected for Lorenz, polarization, and absorption effects were
δ 14.8, 23.3, 29.8, 29.9, 30.1, 32.1, 32.5, 33.3, 130.3, 130.7,
1
31.4, 135.7, 138.3, 140.9, 148.9, 192.6; HRMS (MALDI, DHB):
+
m/z calcd for [C36
Calcd for C36
,4-Dip er flu or ooctyl-2,5-bis(4-(1H-p h en a n th r o[9,10-d ]-
im id a zol-2- yl)p h en yl)ben zen e (6). Compound 4 (0.67 g,
H
46
O
2
+ H] 511.3571, found 511.3592. Anal.
46 2
H O : C, 84.66; H, 9.08. Found: C, 83.59; H, 8.88.
1
J . Org. Chem, Vol. 67, No. 21, 2002 7191

Krebs and Spanggaard
performed using the programs SMART, SAINT,18 and SAD-
1
8
TABLE 6. Da ta fr om P h otoelectr on Sp ectr a for Au
Su bstr a te a n d Com p ou n d s 6 a n d 7
1
9
ABS. Structure solution, refinement of the structures, struc-
ture analysis, and production of crystallographic illustrations
were carried out usingthe programs SHELXS97, SHELXL97,
VB
VAC
compound
EF
EF
cutoff
∆
IP
2
0
20
2
1
and SHELXTL. The structure was checked for higher sym-
6
7
1.60
1.30
2.95
3.25
-45.45
-45.45
-2.35
-2.05
4.55
4.55
2
2
metry and none was found. Crystallographic data (excluding
structure factors) for the structure reported in this paper has
been deposited with the Cambridge Crystallographic Data
Centre as supplementary publication no. CCDC-177458. Cop-
ies of the data can be obtained free of charge on application to
CCDC, 12 Union Road, Cambridge CB2 1EZ, UK (fax: (+44)-
recorded using 50 eV incident photons and a resolution of 0.2
eV. The experimental chamber was equipped with an ion gun,
a mass spectrometer, and a gas dosing system. The sample
holder was electrically isolated from the chamber, and the
sample was kept at a potential of -9.5 V relative to the sur-
rounding instrument. This serves to eliminate the contribution
from the instrument work function at the low energy cutoff.
A clean gold substrate was first entered and sputtered with
ionized argon using an emission current of 22 mA and a poten-
tial of 3 kV. The sputtering was stopped after 1 h; subsequent
measurements of the gold work function gave the reference
value. The samples of 6 and 7 were introduced and the photo-
electron spectra recorded. The onset and cutoff of the intensity
normalized photoelectron spectra were used to compute the
work function in the case of the pure gold substrate and the
ionization potential and valence band edge in the case of the
gold substrates containing the samples. The data are presented
in Table 6.
1
223-336-033; e-mail: deposit@ccdc.cam.ac.uk).
P h otop h ysica l Meth od s. All measurements were per-
formed using Spectrosolv grade solvents, and the solutions
were degassed for 15 min with argon prior to use. Emission
spectra and lifetimes were measured with an instrument
comprised of a 450 W Xe lamp for steady-state measurements
and a nanosecond flashlamp for lifetime measurements. The
detecting system comprises a single photon counting PMT
detector in a peltier cooled housing. All spectra where mea-
sured in a perpendicular geometry using 1-cm quartz cuvettes.
Steady-state mesurements were obtained with a 1.8 nm band-
pass and corrected. The quantum yield was determined using
9
f
,10-diphenylanthracene in cyclohexane (φ ) 1.00, ref 23) as
the fluorescence standard with refractive index and differential
absorption corrections. In all cases lifetimes were single-
exponential and determined using least-squares analysis as
The gold work function, ΦAu, was thus determined to be 5.3
eV. The position of the valence band edge for 6 and 7, EF
2
4
described by Straume. All time-resolved measurements were
conducted with 7.3 nm band-pass.
VB
,
was lower than the Fermi level of gold. The Fermi level in
compounds 6 and 7 are thus situated in the forbidden band
gap of the materials and the distance form the Fermi level of
the samples to the vacuum level is given by the difference
P h otoelectr on Sp ectr oscop y. The samples for ultraviolet
photoelectron spectroscopy measurements (UPS) were pre-
pared by spin-coating a 1 µM solution of 6 and 7 in boiling
chlorobenzene onto a freshly prepared polycrystalline gold
surface. The samples were then dried in a vacuum oven at 50
VB
VAC
between the ionization potential, IP, and EF , giving EF
.
VAC
The vacuum level shift, ∆, is thus obtained as ∆ ) E
-
°
C for 24 h. The photoelectron spectra were recorded at the
F
ΦAu. The vacuum level shift reflects how the molecules orient
ASTRID storage ring at Aarhus University, Denmark. The
beamline consists of an SX-700 monochromator and a hemi-
spherical electron energy analyzer. An ESCA (electron spec-
troscopy for chemical analysis) scan of the samples was
recorded first to check the cleanliness of the area where the
incident photons illuminated the sample and to confirm the
presence of the elements (C, N, and F for 6 and C and N for
at the surface giving a dipole layer. The procedure employed
in the analysis of the data has been reported.25 See also the
Supporting Information.
Ack n ow led gm en t. This work was supported by the
Danish Technical Science Foundation of Denmark
(STVF). We would like to express sincere gratitude to
7
). The photoelectrons were measured at an angle normal to
Ole Hagemann for assistance with the synthetic organic
chemistry, J an Alstrup for the DSC measurements,
Mikkel J ørgensen for helpful discussion and NMR
measurements, and Zheshen Li, Søren V. Hoffmann,
and Philip Hofmann for technical support at the ASTRID
storage ring and at the beamline.
the sam-
ple surface. The ESCA scans were performed with 800 eV pho-
tons and a resolution of 0.4 eV. The photoelectron spectra were
(
18) SMART and SAINT; Area-Detector Control and Integration
Software; Siemens Analytical X-ray Instruments Inc.: Madison, WI,
995.
19) Sheldrick, G. M. Empirical absorption program (SADABS)
written for the Siemens SMART platform.
20) Sheldrick, G. M. SHELX-97; Program for structure solution and
refinement. 1997.
21) Sheldrick, G. M. SHELXTL95; Siemens Analytical X-ray In-
struments Inc.: Madison, WI, 1995.
1
(
Su p p or tin g In for m a tion Ava ila ble: Experimental condi-
tions and analytical data for 8, including assignment of the
NMR signals; excitation spectra for 6-10. DSC data for 6 and
(
7
; an energy level diagram for 6 and 7 based on the photo-
(
electron spectroscopy experiment. This material is available
free of charge via the Internet at http://pubs.acs.org.
(
22) Spek, A. L. Acta Crystallogr. 1990, A46, C-31.
(23) Ware, W. R.; Rothman, W. Chem. Phys. Lett. 1976, 39, 449-
4
53.
J O025592D
(
24) Straume, M.; Fraiser-Cadoret, S. G.; J ohnson, M. L. In Topics
in Fluorescence Spectroscopy; Lakowicz. J . R., Ed.; Plenum Press: New
York, 1991; Vol. 2, pp 177-240.
(25) Salaneck, W. R.; L o¨ gdlund, M.; Fahlman, M.; Greczynski, G.;
Kugler, Th. Mater. Sci. Eng. 2001, R34, 121-146.
7
192 J . Org. Chem., Vol. 67, No. 21, 2002