Simple bisthiocarbonohydrazones as sensitive, selective, colorimetric, and switch-on fluorescent chemosensors for fluoride anions

Article abstract of DOI:10.1002/chem.200600904

Bisthiocarbonohydrazones are found to be a class of sensitive, selective, ratiometric. and colorimetric chemosensors for anions such as fluoride (F ^-) or acetate (Ac^-). The sensitivities, or the binding constants of the sensors with

Full text of DOI:10.1002/chem.200600904

DOI: 10.1002/chem.200600904
Simple Bisthiocarbonohydrazones as Sensitive, Selective, Colorimetric, and
Switch-On Fluorescent Chemosensors for Fluoride Anions
Feng Han,[a] Yuhui Bao,[a] Zhigang Yang,[a] Thomas M. Fyles,[b] Jianzhang Zhao,*[a]
Xiaojun Peng,*[a] Jiangli Fan,[a] Yunkou Wu,[a] and Shiguo Sun[a]
Abstract:
Bisthiocarbonohydrazones
the sensors. Single-crystal structures
and 1H NMR titration analysis shows
switch-on fluorescent chemosensor for
Fꢀ and Acꢀ. The recognition of Fꢀ in
acetonitrile (MeCN) by a sensor with
nitrophenyl substituents is tolerant to
MeOH (MeCN/MeOH=10:1, v/v) and
water (MeCN/H2O=30:1, v/v); at
these solvent ratios the absorption in-
tensity of the sensor–Fꢀ complex solu-
tion at maximal absorption wavelength
was attenuated to half of the original
value in pure MeCN.
are found to be a class of sensitive, se-
lective, ratiometric, and colorimetric
chemosensors for anions such as fluo-
ride (Fꢀ) or acetate (Acꢀ). The sensi-
tivities, or the binding constants of the
sensors with anions, were found to be
strongly dependent on the substituents
appended on the p-conjugation frame-
work, the delocalization bridge CH=N,
the aromatic moiety, and the hetero
atom in the C=X group (X=O, S) of
ꢀ
ꢀ
that the CH=N moiety is a hydro-
gen-bond donor, and it is proposed
that an additional CH···F hydrogen
bond is formed for the sensors in the
presence Fꢀ. A sensor bearing anthra-
cenyl groups is demonstrated as
a
Keywords: anions · colorimetry ·
fluorescence · hydrazones · sensors
Introduction
sensor, hydrogen-bond-induced p-electron delocalization, or
NH deprotonation, are believed to be responsible for signal-
ing the binding event.[16–18] However, Fꢀ chemosensors are
usually synthetically demanding, so the effect of molecular
structure diversity on the performance of the sensor are not
thoroughly investigated. In this article, we report a set of
thiocarbonohydrazones as efficient colorimetric anion sen-
sors that are easy to synthesize.
The amide group (including thioamide) has often been
used as the binding group for anion chemosensors.[16c,17,19,20]
It is known that thioamide protons are stronger hydrogen-
bond donors than the protons in the oxo amide bonds.[21]
The acidity of the (C=S)NH proton with a pKa value of 11–
13, is much higher than an oxo amide with a pKa of 17.[21a]
Considering that the thiourea group offers an ideal binding
and signal transduction unit,[20] we envisage utilizing this
property with a 1,3-diaminothiourea, which contains two
free amino groups for condensation with various aldehydes
or ketones, so that the structure of the sensor molecules can
be easily modified. Our results show that the class of bis-
thiocarbonohydrazone compounds are easily prepared, and
are sensitive, selective, ratiometric, and colorimetric chemo-
sensors for anions such as Fꢀ and Acꢀ. The recognition
system is tolerant to protic solvents such as H2O and MeOH
in mixed MeCN solvents. The UV/Vis absorption spectra,
1H NMR titration, and single-crystal structures all suggest
Much attention has been paid to the development of chemo-
sensors for anions such as fluoride (Fꢀ) or acetate (Acꢀ),
due to the important role these anions play in the environ-
mental, security, and health sciences. Many synthetic recep-
tors have been reported, in which various binding and sens-
ing mechanisms are employed,[1–8] such as hydrogen bond-
ing,[9–12] binding-induced perturbation of the p-conjugation
framework of the sensor molecules,[13,14] or the catalysis
effect of Fꢀ to trigger a chemical reaction.[15] For the sensors
based on hydrogen bonds, urea or thiourea units, among
others, are usually used as the binding sites. For this kind of
[a] F. Han, Y. Bao, Z. Yang, Prof. J. Zhao, Prof. X. Peng, Dr. J. Fan,
Y. Wu, Dr. S. Sun
State Key Laboratory of Fine Chemicals
Dalian University of Technology
158 Zhongshan Road, Dalian 116012 (P.R. China)
Fax : (+86)411-3960-8007
E-mail: zhaojzh@dlut.edu.cn
pengxj@dlut.edu.cn
[b] Prof. T. M. Fyles
Department of Chemistry, University of Victoria
Victoria, BC, V8W3V6 (Canada)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
2880
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2880 –2892

FULL PAPER
that the CH=N proton is involved in the binding with Fꢀ
through a CH···F hydrogen bond.
disappeared. The stepwise UV/Vis spectral changes corre-
spond to the stepwise deprotonation of the two NH protons,
and we suggest that a fully conjugated system was formed
that is responsible for the absorption at 605 nm.[16c,17] Sensor
1 can be described as a ratiometric sensor. This is also the
case for most of the other sensors described in this paper.
Such a sensitive and drastic (240 nm red-shifted) change in
the absorption is ideal for a colorimetric sensor.
Results and Discussion
Synthesis: The thiocarbonohydrazones (1–3, 5–7, 9–12) were
prepared by condensation of aldehydes or ketones with car-
bohydrazides or thiocarbohydrazides. Two commercially
available compounds (4, 8) were used as controls. Electron-
donor as well as electron-acceptor substituents, and different
aromatic moieties were incorporated into the chemosensor
to tune its sensing performance, as the acidity or the hydro-
The deprotonation of 1 by Fꢀ is a relatively slow process.
If the titration is carried out with small aliquots of Fꢀ, then
the overall titration time is sufficient to result in the tailing
absorption at an early stage of the titration (Figure 1a).
The significant changes of the UV/Vis spectra of sensor 1
upon titration with Fꢀ is attributed to the deprotonation of
the thioamide protons with a three-step process, similar to
the sequence previously reported.[17,18] This was confirmed
ꢀ
ꢀ
gen-bonding ability of the NHCAHTRE(UNG C=X)NH framework (X=
O, S), and the absorption/emission maxima (in aspects of
both wavelength and intensity) is altered by variation of the
substituent groups.
1
by the H NMR titration (see below). With the addition of
five equivalents of Fꢀ to
a solution of sensor 1 in
Effect of electron push/pull groups on the recognition:
Among the sensors illustrated in Scheme 1, compound 1 is
expected to show the tightest binding with Fꢀ, due to the en-
hanced acidity of the thioamide protons and the stabiliza-
[D6]DMSO, the NH peaks disappeared, whilst the peaks of
the FHFꢀ species were observed at d=16.13 ppm.[24]
In the UV/Vis spectra, the first changes at low amounts of
Fꢀ added are due to a hydrogen-bonded complex. This spe-
cies is responsible for the new absorbance at 407 nm. As the
amount of Fꢀ increases, the sensor is deprotonated to the
conjugate base with formation of the FHFꢀ ion. The conju-
gate base of 1 is responsible for the absorption at 495 nm.
At even higher Fꢀ concentration a second deprotonation
occurs to produce a second FHFꢀ ion and the dianion of 1,
which shows the absorption at 605 nm. The three equilibria
are given in Equations (1)–(3).
tion effect of the negative deprotonated species induced by
[20]
ꢀ
the electron withdrawing nitro ( NO2) group. With pro-
gressive addition of Fꢀ (Figure 1), the UV/Vis absorption of
1 at 360 nm was diminished, whilst new peaks at 407 and
495 nm appeared. The absorption at 360 nm is mainly due to
[22,23]
ꢀ
ꢀ
the Ar CH=N NH conjugation framework.
The well-
located isosbestic points at 297 and 397 nm indicated the
shifting of a well-defined binding equilibrium in the solution
by addition of Fꢀ (Figure 1a). After more than four equiva-
lents of Fꢀ was added, however, the tailing absorption at
600 nm appeared and was intensified, with a new isosbestic
point at 535 nm. This set of changes was clearly demonstrat-
ed with a “quick” titration of 1 with Fꢀ (Figure 1b). With
400 equivalents of Fꢀ added, the absorption at 605 nm
reached its maximum, whilst the initial absorption at 360 nm
LH2 þ Fꢀ Ð LH2 Á Á Á F KH
ð1Þ
ð2Þ
ð3Þ
LH2 Á Á Á F þ Fꢀ Ð LHꢀ þ FHFꢀ K1
LHꢀ þ 2Fꢀ Ð L2ꢀ þ FHFꢀ K2
Scheme 1. Structures of the chemosensors used in this study.
Chem. Eur. J. 2007, 13, 2880 –2892
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2881

J. Zhao, X. Peng et al.
gave logKapp =5.11 and logK2 =4.84 for the deprotonation
process. A more complete analysis used data at three wave-
lengths (360, 412, and 598 nm) to fit the data simultaneously
to the model of Equations (1)–(3) to give logKH =4.64ꢁ
0.04, logK1 =4.39ꢁ0.03, and logK2 =5.40ꢁ0.14 (Figure 1c).
ꢀ
ꢀ
The binding of 1 with Acꢀ, HSO4 , and H2PO4 were also
studied. The spectral changes of 1 upon titration with differ-
ent anions are shown in Figure 2. It was found that besides
Figure 2. UV/Vis absorption changes at 487 nm for sensor 1 (1.0”10ꢀ5
m
),
in MeCN, 208C) with addition of selected anions. Fꢀ (&), Acꢀ
(
*
ꢀ
ꢀ
~
^
( ), and HSO4 ( ).
H2PO4
Fꢀ, 1 shows a high sensitivity toward Acꢀ. The wavelength
of this experiment (487 nm) detects the formation of the
conjugate base of 1 according to Equation (4)
LH2 þ Xꢀ Ð LHꢀ þ XH KX
ð4Þ
The titration with Acꢀ gave a value of logK=5.10ꢁ0.03.
ꢀ
Sensor 1 showed no response to HSO4 , in contrast to previ-
ous reports of the oxo analogues.[18] Considering that no ad-
ditional discrimination mechanism is working for such a
simply structured sensor, 1 shows a relatively good selectivi-
ꢀ [25]
ty toward Acꢀ over H2PO4 .
Furthermore, the sensor
demonstrated perfect selective recognition of two tetrahe-
dral anions, HSO4 and H2PO4 . For H2PO4 , a binding con-
Figure 1. UV/Vis spectral changes of 1 (1.0”10ꢀ5 m in MeCN) upon addi-
tion of Bu4NF, 208C. a) From 0 to 7 equivalents [Bu4NF]: Inset: absorb-
ance at 487 nm versus equivalents of Fꢀ added. b) Selected key curves of
the quick titration of sensor 1 with [Bu4NF]. c) Simultaneous fitting of
the titration data at three wavelengths to a three-step model of hydrogen
bonding, first deprotonation, and second deprotonation (see text).
ꢀ
ꢀ
ꢀ
stant of logKX =4.13ꢁ0.24 was determined, whereas the
ꢀ
sensor gives no response to HSO4 .
ꢀ
Without the electron withdrawing NO2 group on the
phenyl ring, that is, sensor 2, a much smaller red shift of
71 nm (from 322 to 393 nm, Figure 3) of the UV/Vis absorp-
tion was observed upon addition of Fꢀ.
The spectral changes due to the formation of the hydro-
gen-bonded complex and the first deprotonation occur in
the same concentration range of added Fꢀ so the individual
steps cannot be analyzed in isolation. Our analysis was done
on three different levels. If the signal at a given wavelength
appeared to plateau with a small amount of added Fꢀ
(<10 equiv), then the binding could be treated as a simple
1:1 binding isotherm in which KH and K1 were combined
into an apparent K. The inset of Figure 1a shows such an
analysis to give a value of logKapp =4.72ꢁ0.03. Using this
value as a fixed input, a value for logK2 =4.02ꢁ0.03 was de-
termined for the second deprotonation. These are only ap-
parent values: titration of 1 with Fꢀ carried out with a lower
sensor concentration of 2.0”10ꢀ6 m,[18] analyzed this way
From the UV/Vis absorption changes of 2 (Figure 3), it
appeared that the deprotonation of the NH group occurred
1
upon addition of Fꢀ, which is also proved by the H NMR ti-
tration of sensor 2 with Fꢀ in [D6]DMSO in which the FHFꢀ
peaks appeared at 16.1 ppm. However, in contrast to sensor
1, we postulate that for sensor 2, only one thioamide proton
can be removed by Fꢀ.[17]
An apparent 1:1 binding constant of logKX =3.65ꢁ0.06
was determined for sensor 2 with Fꢀ, comparable to a re-
ported Fꢀ sensor containing a hydrazone moiety.[16a] Sensor
1 gives a much higher binding constant for Fꢀ. This is consis-
tent with the report that an electron-withdrawing group
would improve the binding of amide with anions.[25] The
2882
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2880 –2892

Hydrazones as Fluoride Chemosensors
FULL PAPER
Figure 3. UV/Vis spectral changes of sensor 2 (3.3”10ꢀ5 m in MeCN)
upon addition of Bu4NF, at 208C: Inset: absorbance at 393 nm versus
equiv of Fꢀ added.
Figure 5. UV/Vis spectral changes of sensor 3 (1.0”10ꢀ5 m in MeCN)
upon addition of Bu4NF, at 208C: Inset: absorbance at 460 nm versus
equivalents of Fꢀ added.
binding of 2 with several anions was studied, and the bind-
ing isotherms are illustrated in Figure 4.
An apparent 1:1 binding constant of logKX =3.08ꢁ0.03
was determined for the recognition of Fꢀ by 3. This value is
lower than that of 1 or 2. With HSO4 , a binding constant of
ꢀ
logK=3.47ꢁ0.15 was determined.
The analogue 4 was reported to display a monotonic in-
crease of the UV/Vis absorption maximum at 274 nm, with
a slight red shift to 278 nm in presence of Acꢀ,[22] due to hy-
drogen bonding. With Acꢀ, a binding constant was deter-
mined as logK=4.46ꢁ0.06 for 4, which is in good agree-
ment with the reported value of logK=4.20.[22a] Incorpora-
ꢀ
ꢀ
tion of the C=N moiety into the sensor leads to 2, for
which a red shift of 70 nm, from 322 to 393 nm, was ob-
served in the presence of Acꢀ.
It has been reported that a urea derivative can undergo
deprotonation in the presence of Fꢀ;[16c,18] therefore we rein-
vestigated the titration of 4 with Fꢀ. The UV/Vis absorption
at 274 nm was only slightly intensified with the addition of
up to seven equivalents of Fꢀ. With more than seven equiva-
lents of Fꢀ added, however, a new peak at 358 nm appeared.
This new peak is attributed to the deprotonation of NH.
The stability constant for the hydrogen bonded, and there-
after the deprotonated, species were determined as logKH =
4.33ꢁ0.14 and logK1 =3.35ꢁ0.08, respectively.
Figure 4. UV/Vis absorption changes of 2 (MeCN, 208C) at 393 nm with
ꢀ
ꢀ
addition of anions Fꢀ(&), Acꢀ
(
), H2PO4
(
), and HSO4
(
~
*
^
).
For 2, a much lower binding constant with Acꢀ was ob-
served (logKX =3.12ꢁ0.06) compared to that of sensor 1
(logKX =5.10ꢁ0.03), but a greatly improved selectivity
toward Acꢀ over H2PO4 was observed. Sensor 2 shows a
ꢀ
ꢀ
ꢀ
negligible response to H2PO4 and HSO4 . The binding of 2
with anions in DMSO was also studied and a similar result
was observed.
Our results show that introduction of the CH=N benzyli-
dine group into the sensor is important, as a 70 nm red-shift
of the UV/Vis absorption is produced in the presence of Fꢀ,
and the deprotonation of the thioamide becomes easier.
The selectivity of the sensors toward different anions is
partially due to the intrinsic pKa and geometry of the
anions. However, it is also possible to design a structurally
more complicated sensor that will be selective for Fꢀ; for
example, a sensor with a cage structure for which the size
selectivity toward specific anions can be employed to im-
prove selectivity.[26]
Electron-donating substituents on the phenyl ring will de-
crease the acidity of the thioamide proton, as well as the sta-
bilization efficiency of the negatively charged deprotonated
species. This is illustrated by sensor 3, with a dimethylamino
group at the para-position of the phenyl ring. The UV/Vis
absorption spectral profile of 3 with addition of Fꢀ is depict-
ed in Figure 5; with the addition of up to 300 equivalents of
Fꢀ only one peak at 460 nm developed.
Mono- or bisthiocarbonohydrazone: To elucidate whether
both branches of the symmetric bisthiocarbonohydrazones
were essential for recognition of anions, monosubstituted
asymmetric thiourea compounds 5 and 6, which have only
one hydrazone branch, were used as controls. For 5, the ab-
sorption maxima at 358 nm first made a minor red-shift to
370 nm upon addition of Fꢀ (up to 100 equiv Fꢀ), then a
blue-shifted peak (instead of a red-shifted peak for 1) at
342 nm developed. The spectral change of 6 shows that the
absorption peak at 356 nm was diminished with addition of
Fꢀ, but only a very weak absorption peak developed at
Chem. Eur. J. 2007, 13, 2880 –2892
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2883

J. Zhao, X. Peng et al.
455 nm. This indicates an inefficient deprotonation, or at
least a poor signal transduction efficiency for the binding
event. Instead, the bisthiocarbonohydrazones 1 and 2 give
much more significant spectral responses to anions. There-
fore, a bis-substituted thiocarbonohydrazone is essential for
employing these compounds as Fꢀ sensors. The interactions
ꢀ
ꢀ
of sensors 5 and 6 with Acꢀ, HSO4 , and H2PO4 were also
studied, and the results show that the response profiles are
similar to those of bis-substituted sensors, such as 1 and 3.
The effect of CH=N on the binding: For sensor 2, only a de-
protonation process was observed in presence of Fꢀ. For
sensor 4, a clear stepwise process of hydrogen bonding and
then a deprotonation was observed, possibly due to the in-
creased pKa of the NH proton, without the conjugation
ꢀ
ꢀ
moiety of CH=N . Thus, we decided to carry out further
ꢀ
ꢀ
studies of the role of the CH=N moiety in recognition.
As the single-crystal structure and 1H NMR titration
infers that CH=N may take part in the binding of Fꢀ, we
substituted the CH=N proton with a methyl group to elimi-
nate the possible supporting binding effect of the CH=N
protons. Thus 7 was synthesized. The titration of 7 with Fꢀ
shows that with three equivalents of Fꢀ, the UV/Vis absorp-
tion maximum at 353 nm diminished, whilst a new peak at
491 nm developed, with isosbestic points at 295 and 388 nm
(Figure 6). With more Fꢀ added, a new absorption at 620 nm
appeared, with an isosbestic point at 512 nm. This clear step-
wise change is due to the distinct one-by-one deprotonation
of the two NH protons. Unlike sensor 1, there is no distinct
spectral change associated with the formation of the hydro-
gen-bonded complex for sensor 7. The titration data for 7
can be fitted using the same three-step model used above to
give values of logKH =2.82ꢁ0.2, logK1 =7.64ꢁ0.05 and
logK2 =6.96ꢁ0.04. The value for logKH is substantially
ꢀ
ꢀ
lower for 7 than for 1, indicating that the CH=N proton
plays a role in the stabilization of the hydrogen-bonded
complex in 1. Conversely, sensor 7 is a stronger acid than
sensor 1. The first complete deprotonation (LH2 +2Fꢀ gives
LHꢀ +FHFꢀ) for sensor 1 has a logK=4.64+4.39=9.0ꢁ
0.1, while logK=10.5ꢁ0.1 for 7 makes sensor 7 about 30
times more acidic. Similarly the second deprotonation for 7
is 30 times greater than the second deprotonation of 1. The
electron-withdrawing nitrophenyl substituents are the same
in both cases, so the difference must lie in conformational or
solvation differences as a result of the methyl or hydrogen
substitution between the two compounds.
Figure 6. UV/Vis spectral changes of sensor 7 (1.0”10ꢀ5 m in MeCN)
upon addition of Bu4NF, at 208C: a) from 0 to 3 equivalents Fꢀ, inset: ab-
sorbance at 491 nm versus equivalents of Fꢀ added; b) from 3.5 to
300 equivalents of Fꢀ, inset: absorbance at 491 nm versus equivalents of
Fꢀ added. c) Simultaneous fitting of the titration data at three wave-
lengths to a three-step model of hydrogen bonding, first deprotonation,
and second deprotonation (see text).
The differences between sensors 1 and 7 are best illustrat-
ed in the species distribution diagrams calculated from the
determined values as illustrated in Figure 7. It is evident
that at the sensor concentration of 1.0”10ꢀ5 m the hydrogen
bonded complex makes a significant contribution to the spe-
ciation for sensor 1, but it is negligible for sensor 7. As a
result, there are four significant species in a Fꢀ/1 mixture
but only three in a Fꢀ/7 mixture. This has a direct conse-
quence in the use of 1 and 7 as naked-eye sensors for Fꢀ
(see below). The shift of the mono- and dianionic forms of 7
relative to the same species formed by 1 is a direct result of
the higher acidity of 7 relative to 1, since the base (Fꢀ) is a
constant factor.
ꢀ
ꢀ
To elucidate the role of the CH=N group as a conjuga-
tion bridge for the sensing of anions, sensor 8 was investigat-
ed. The UV/Vis absorption of sensor 8 with addition of Fꢀ
demonstrated only a minor variation. This means that the
ꢀ
ꢀ
CH=N moiety also contributes to the recognition ability
of the sensors, possibly through which an intramolecular
charge transfer is facilitated or the deprotonated species is
effectively stabilized.
2884
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2880 –2892

Hydrazones as Fluoride Chemosensors
FULL PAPER
The only possibility is that a CH···F hydrogen bond is
formed. The binding constant of sensor 9 with Fꢀ was deter-
mined as logKH =2.36ꢁ0.06, a low value compared to the
sensors with CH=N groups, such as 1–4. However, it is
known that the cooperative effect of a second binding site
will produce a much higher binding constant than a mono-
dentate sensor.[30] The possibility of the strong CH···F inter-
1
action was further investigated by H NMR titration of the
sensor 9 with Fꢀ (see below).
To clarify that the UV/Vis spectral variation in Figure 8
was exclusively due to the CH···F interaction (CH=N in 9 as
hydrogen-bond donor), sensor 10 was synthesized as a con-
trol, in which the CH=N hydrogen is substituted with a
methyl group. The titration of sensor 10 with Fꢀ showed
that the UV/Vis spectra did not change at all. This result
gives clear and direct evidence that an interaction occurred
between the CH=N hydrogen and Fꢀ for sensor 9. It is an-
ticipated that the same interactions exist for other sensors,
especially 1, 6, and 11.
Thiocarbonohydrazone or carbonohydrazone—the effect of
one-atom variation on recognition: For this purpose, carbon-
ohydrazone 11 was synthesized. The titration of 11 with Fꢀ
showed that only one new UV/Vis absorption peak at
497 nm appeared, whilst the original UV/Vis absorption at
344 nm diminished. Compound 11 failed to show an absorp-
tion at 600 nm even with 300 equivalents of Fꢀ. This result
indicated a lower acidity of the amide protons and that only
one deprotonation occurred for 11. The binding constant
was determined as logKX =4.31ꢁ0.04, only 33% of that for
sensor 1.
Figure 7. Distribution diagram of the species for a) sensor 1 and b) sensor
7 titrated with [Bu4N]F (1.0”10ꢀ5 m in MeCN) at 208C.
As 1 and 7 showed significantly different binding perform-
ances to Fꢀ, and that a CH···F hydrogen bond has been re-
ported for some anion chemosensors,[8k,27–29] we suspected
ꢀ
ꢀ
that the CH=N proton of the present thiocarbonohydra-
zones may also be involved in the binding of anions, espe-
cially Fꢀ. Therefore, sensor 9, in which the NH proton is
missing, was synthesized as a negative control. No UV/Vis
absorption changes were initially expected to occur for 9 in
the presence of Fꢀ, because the essential Fꢀ binding site, the
NH moiety, is missing and the CH···F interaction is expected
to be quite weak.
In the presence of Acꢀ, 11 shows a red-shifted UV/Vis ab-
sorption from 344 to 383 nm with a isosbestic point at
359 nm. A similar spectral response was observed in the
ꢀ
ꢀ
presence of H2PO4 , but no response to HSO4 was ob-
served. This performance is totally different from 1, which
showed a UV/Vis absorption at 490 and 605 nm with Acꢀ.
The minor red-shift of 11 in the presence of Acꢀ and
ꢀ
To our surprise, however, a red-shifted UV/Vis absorption
H2PO4 is probably due to hydrogen bonding, not to depro-
was observed for sensor 9 with addition of Fꢀ (Figure 8).
tonation of the NH groups.[18]
This result shows that the one-atom transition from
oxygen (amide) to sulfur (thioamide) is very important for
the performance of the present carbonohydrazone chemo-
sensors. A greatly improved sensing performance toward Fꢀ
was observed with the thiourea 1 compared to the urea 11,
two sensors with only one atom variation.
A chromophore of an anthracene moiety was also incor-
porated into the carbonohydrazone (sensor 12), but no sig-
nificant red-shifted UV/Vis absorption was observed in the
1
presence of Fꢀ, and the H NMR results showed that depro-
tonation did not occur (at least with 10 equiv Fꢀ).
The binding constants of the sensors with various anions
are summarized in Table 1. It was found that the selectivity
of the sensors toward anions can be drastically modified
simply by changing the substitution groups on the phenyl
moiety. For example, sensor 3 is highly selective toward
Figure 8. UV/Vis spectral changes of sensor 9 (1.0”10ꢀ5 m in MeCN)
upon addition of Bu4NF, at 208C. Inset: absorbance at 430 nm versus
equivalents of Fꢀ added.
ꢀ
ꢀ
HSO4 over H2PO4 , with a binding constant of logK=
Chem. Eur. J. 2007, 13, 2880 –2892
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2885

J. Zhao, X. Peng et al.
Table 1. Binding constants and absorption maxima of the sensors with anions in MeCN solution, 208C.[a]
Then we carried out the ti-
tration of 1 with Fꢀ in MeCN/
MeOH mixed solvent (30:1,
v/v). Note the molar concen-
tration of MeOH in the mixed
solvent is about 0.79m, nearly
79000 times the concentration
of 1. The result shows that
even with 1000 equivalents of
Fꢀ, only one new UV/Vis ab-
sorption peak at 474 nm ap-
peared, indicating only one de-
protonation of the two NH
groups of 1 (Figure 9b). The
binding constant was deter-
mined as logKX =3.51ꢁ0.01.
This is a 40-fold decrease from
the value in pure MeCN
(logK=5.11). We notice the
discrepancy between Figure-
s 9a and 9b; more work is
needed to elucidate the com-
plexities of these mixed sol-
Sensor
LogKX with anions
Absorption maxima [nm]
ꢀ
ꢀ
Fꢀ
CH3COOꢀ
HSO4
H2PO4
no Fꢀ
with Fꢀ
[e]
1
4.64ꢁ0.04[b]
4.39ꢁ0.03[c]
5.40ꢁ0.14[d]
3.65ꢁ0.06
5.10ꢁ0.03
–
4.13ꢁ0.24
360
491
605
[e]
[e]
2
3
4
3.12ꢁ0.06
–
–
–
322
376
274
274
358
356
353
393
460
274
358
342
455
491
620
[e]
[e]
3.08ꢁ0.03
–
3.47ꢁ0.15
5.08ꢁ0.10[d]
4.10ꢁ0.09[b]
2.94ꢁ0.02
4.33ꢁ0.14[b]
3.35ꢁ0.08[c]
4.19ꢁ0.07
4.46ꢁ0.06
3.98ꢁ0.03[d]
[e]
5
6
7
2.30ꢁ0.13
4.23ꢁ0.00
4.80ꢁ0.04
–
[e]
4.58ꢁ0.08
–
–
3.71ꢁ0.03
[e]
2.80ꢁ0.20[b]
7.60ꢁ0.05[c]
6.96ꢁ0.04[d]
4.46ꢁ0.03
[e]
[e]
8
9
10
11
12
–
–[e]
3.68ꢁ0.09
–
–
–
283
325
321
344
404
345
430
321
497
500
[e]
[e]
[e]
2.36ꢁ0.06
–
–
–
–
–
–
[e]
[e]
[e]
[e]
–
[e]
[e]
4.31ꢁ0.04
4.63ꢁ0.04
4.13ꢁ0.02
[e]
4.36ꢁ0.04
3.79ꢁ0.05
–
[a] 1.0”10ꢀ5 m of sensor in acetonitrile. Constants determined by fitting a 1:1 binding model. For most cases,
the determination coefficients r2 >0.98. [b] logKH. [c] logK1. [d] logK2. [e] No spectral changes were observed
or the minor change was not suitable for accurate measurement of the KX values.
ꢀ
ꢀ
3.47ꢁ0.15 for HSO4 and no response for H2PO4 . With
sensor 1, however, the selectivity was totally reversed, that
vent, multiple-deprotonation equilibria.
The resistance to H2O and MeOH of the recognition of
Fꢀ by sensor 1 is possibly due to the high acidity of the NH
protons. This is promising and we envisage that a chemosen-
sor for Fꢀ, which is applicable in aqueous solution, might be
developed with thiocarbonohydrazone structural units.
ꢀ
is, no response to HSO4 , but a binding constant of logKX =
ꢀ
4.13ꢁ0.24 was observed for H2PO4 . This is interesting,
ꢀ
ꢀ
since in principle, the concentrations of H2PO4 and HSO4 ,
both biologically significant anions, could be independently
determined in a mixture containing both ions.
MeOH and H2O tolerance of the recognition: For the hy-
drogen-bonding-based Fꢀ chemosensors, except for a few
successful cases,[8l] only aprotic solvents are usually permit-
ted. In protic solvents such as H2O or MeOH, the binding of
the sensor with anions would be completely quenched due
to the effective competition of the solvent mole-
cules.[10,12,18,22] As the pKa of thioamide protons is in the
range of 11–13,[21a] and the pKa values of H2O and MeOH
are 15.7 and 15.5, respectively, we speculated that the thio-
carbonohydrazone Fꢀ sensor would be tolerant to H2O and
MeOH, at least to some extent.
Initially, we carried out a re-titration of the deprotonated
1 (in MeCN) with pure H2O and MeOH. Firstly the absorp-
tion of 1 was brought from 361 nm to its maximum of
605 nm by addition of 400 equivalents of Fꢀ, and then the
solution was titrated with pure MeOH. The result shows
that the recognition system is tolerant to MeOH and H2O
(the result with MeOH is depicted in Figure 9a; a similar
result was obtained with H2O). With a solvent ratio of
MeCN/MeOH=10:1 (v/v) or MeCN/H2O=30:1 (v/v), the
absorption at 605 nm was attenuated to half of the original
values in pure MeCN. It must be noted that the volume of
the solution changed significantly during the re-titration
with MeOH (in total 1.0 mL of MeOH was added to 3.0 mL
of the sensor 1–Fꢀ solution), therefore no well-located iso-
sbestic point was observed.
Figure 9. MeOH tolerance of recognition of Fꢀ by 1. a) First, 400 equiva-
lents Fꢀ was added to the solution of 1 in MeCN, then the solution was ti-
trated with pure MeOH; b) UV/Vis spectral changes of 1 (1.0”10ꢀ5 m in
MeOH/MeOH (30:1 V/V) at 208C) upon addition of Bu4NF.
2886
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2880 –2892

Hydrazones as Fluoride Chemosensors
FULL PAPER
Crystal structure and the molecular conformations: The
single crystal of sensor 12 revealed a non-coplanar confor-
mation. Intermolecular hydrogen bonds, including sensor–
sensor and sensor–THF, as well as intermolecular face-to-
face p–p interactions were found for the packing of the mol-
ecules in the crystal (Figure 10).
side arms is highly likely. The UV/Vis absorption changes
with addition of Fꢀ supports this assumption.
It was found that hydrogen bonding and p–p interactions
are responsible for the packing of the sensor molecules
within the single crystal, which is different from that report-
ed for thiocarbonohydrazones.[16,31] The intermolecular hy-
drogen bond length S(1)···HACHTER(UNG 4 A) is about 2.517 Š, which is
shorter than that of the reported analogues,[16,31] inferring a
stronger hydrogen bond.
It is worth noting that a hydrogen bond is formed be-
ꢀ
ꢀ
tween the CH=N proton and the oxygen atom from the
solvent molecule of THF in the single crystal. The geome-
tries of the hydrogen bonds are summarized in Table 2. The
Table 2. Geometrical features of the CH···O interactions in the single
crystal of sensors 12 and 2.
ꢀ
Donor
group
D···A
[Š]
H···A
[Š]
D H···A
[8]
acceptor
atom
ꢀ
12
2
N(2) H
N
3.103
3.166
3.364
3.506
3.723
2.345
2.341
2.517
2.670
2.884
147.07
147.60
168.65
158.84
148.02
O(1)
O(1)
S(1)
S(1)
S(1)
ꢀ
C(17) H
ꢀ
N(4) H
ꢀ
N(2) H(N2)
ꢀ
C(14) H(14)
CH···O interactions noted are similar to other hydrogen
bonds of this type.[33,34] The bond lengths for O(1)···H
(17 A)
and O(1)···H(2 A) are 2.341 and 2.345 Š, respectively. Fur-
ACHTREUNG
AHCTREUNG
thermore, the hydrogen bond length of D···A is 3.166 Š,
which is close to the shortest bond length reported
(3.135 Š) for this type of CH···O hydrogen bond.[33b] This
short bond length, together with the fact that the two hydro-
gen bonds O(1)···HACHTRE(UNG 17A) and O(1)···H(2A), with CH=N
and NH as the donors, respectively, give almost the same
bond length, inferring that the CH=N is a strong hydrogen
bond donor, and it is highly possible for this proton to form
a hydrogen bond with Fꢀ in nonprotic solution.[8k,33] This hy-
Figure 10. Single-crystal structure of sensor 12. Top: Hydrogen-bond
motif with THF. Bottom: Molecular packing mode showing the p–p in-
teraction of 12.
1
drogen bond formation is also supported by the H NMR ti-
The dihedral angle of the two anthracene moieties of 12 is
tration study of the sensor with progressive addition of Fꢀ
(see below).
ꢀ
ꢀ
about 738. The torsion angles between the two CH=N
groups and the planes of the proximate anthracene moieties
The single-crystal structure of compound 2 is different
from that of compound 12. There is no face-to-face p–p in-
teraction or solvent molecules in the crystal of 2. It is worth
noting that the two phenyl rings are nearly coplanar, in con-
trast to compound 12. The torsion angles between the
phenyl moieties and the thiourea groups are 177.63 and
13.388. This profile infers that after deprotonation a fully
conjugated molecular framework is possible in the presence
of Fꢀ. Intermolecular hydrogen bonding was observed for 2,
with the NH and the CH=N moiety as the hydrogen bond
donors. The bond lengths (H···S) for NH···S and CH···S are
2.670 and 2.884 Š, respectively (Table 2). The bond length
of CH···S (2.884 Š) for the intermolecular hydrogen bond
also supports that the CH=N moiety can act as a hydrogen-
bond donor.[16a]
ꢀ
are 173.738 and 119.448, respectively. This means one CH=
ꢀ
N
group is preorganized to be nearly coplanar with the an-
ꢀ
ꢀ
thracene moiety to its proximity, but another CH=N
group takes a perpendicular orientation to its proximate an-
thracene moiety. For the former, the C=N bond length
ꢀ
N(1) C(17) is 1.250 Š, shorter than the other one, which is
not coplanar with the anthracene, which has a bond length
ꢀ
ꢀ
N(3) C(15) of 1.266 Š. This result infers that for N(1)
C(17), there exists a conjugation with the anthracene
moiety.[16,31]
The two anthracene moieties take such a geometry that
both participate in the off-set face-to-face p–p interaction
with the anthracene moiety from another sensor molecule.
In this interaction the distance between the aromatic rings is
about 3.4 Š. This short distance indicates a strong p–p inter-
action.[32] The single-crystal structure shows that delocaliza-
tion on one of the preorganized coplanar carbonohydrazone
Hydrogen atoms of aromatic rings or methylene groups
have been reported to be able to form a CH···X type of hy-
drogen bond with anions such as Fꢀ,[8k,35] and have been de-
Chem. Eur. J. 2007, 13, 2880 –2892
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2887

J. Zhao, X. Peng et al.
tected by 1H NMR titration studies of the sensor with
spectra. This is consistent with our earlier studies that
showed that thioamide is characterized by much higher rota-
tional barriers than its oxo counterparts.[23,37]
anions.[36] To our knowledge, no Ar CH=N has been re-
ꢀ
ꢀ
ported as a hydrogen bond donor in chemosensors.
1
The H NMR titration of sensor 1 shows that upon addi-
NMR titration and the sensing mechanism: Both the UV/
Vis absorption of sensor 9 and especially the single-crystal
structure of sensor 12 present a clue that a CH···F hydrogen
bond is possible for the present sensors. Therefore, a
1H NMR titration study of sensor 1 with progressive addi-
tion of Fꢀ was carried out (Figure 11).
tion of only 0.1 equivalents of Fꢀ, the two peaks of the
amide proton show severe broadening and collapse to a
single peak. An unanticipated dramatic merging and broad-
ꢀ
ꢀ
ening of the CH=N proton peaks occurred and a new,
very broad peak appeared. With addition of up to 0.5 equiv-
alents of Fꢀ, the NH peak disappeared completely and the
ꢀ
ꢀ
AHCTREUNG CH=N peak recovered and moved upfield. At the same
time the peak at d=8.02 ppm moved initially downfield,
and then made an upfield return. This change was accompa-
nied by the disappearance of the amide proton and the re-
ꢀ
ꢀ
covery of the CH=N proton peak.
We attribute this series of changes to the first two steps of
the Fꢀ recognition by sensor 1: first, formation of a hydro-
gen bond CH···F at very low Fꢀ concentration, and second,
deprotonation of the amide bond with progressive addition
of the Fꢀ.
With addition of 0.1 equivalents Fꢀ, a CH···F hydrogen
ꢀ
ꢀ
bond was formed, indicated by the merging of the CH=N
peaks and the broadness of the new peak.[8k,28,38] The single-
ꢀ
ꢀ
crystal structure of 12 suggests that the CH=N proton is a
hydrogen-bond donor (Figure 9). It was reported that Fꢀ
can form a cyclic hydrogen bond with thioamide pro-
tons.[39,40] The CH···F hydrogen bond has also been reported
for anion chemosensors, for example, Fꢀ can form a hydro-
gen bond with the hydrogen of a benzene ring,[41] a methyl-
ene group,[42] or a hydrogen atom from an anthracene
group,[14] or (C=O)CH=C(OH).[8k] To our knowledge, no
ꢀ
ꢀ
CH···F hydrogen bond in which Ar CH=N acts as the hy-
drogen-bond donor has been reported.
With addition of 0.1 equivalents Fꢀ to a solution of 1, the
peak of Hc at d=8.07 ppm made a progressive downfield
shift, due to the space effect of the hydrogen bond (polariza-
ꢀ
tion of the C H bond), whilst for the Hd proton, the
through-bond effect is dominant and a upfield shift was ob-
served.[18]
With more Fꢀ added (0.5 equiv and more), the NH pro-
tons were significantly deprotonated, so the NH peak disap-
Figure 11. 1H NMR (400 MHz) spectra of sensor 1 (5 mm) in [D6]DMSO.
a) Sensor 1 only; b) 1 + 0.1 equivalents Fꢀ; c) 1 + 0.2 equivalents Fꢀ;
d) 1 + 0.3 equivalents Fꢀ; e) 1 + 0.5 equivalents Fꢀ; f) 1 + 0.7 equiva-
lents Fꢀ; g) 1 + 1.0 equivalent Fꢀ, inset: partial 1H NMR with 5 equiva-
lents Fꢀ.
ꢀ
ꢀ
peared. The CH=N proton itself is not a strong enough
hydrogen-bond donor to independently hold a Fꢀ tightly
(proved with the binding constant of sensor 9 with Fꢀ, which
is only 0.5% of that for sensor 1), therefore, without the co-
binding effect of the proximate NH proton, no Fꢀ is bonded
ꢀ
ꢀ
tightly by the CH=N moiety, and its proton peak is recov-
ered. Due to the partial deprotonation of the thioamide, the
electron density of the thioamide framework will increase
It is known that the conformations of thiourea derivatives
are characterized by a restricted rotation in solution.[21b] The
1H NMR spectrum of sensor 1 shows that the two NH
and the peak of the CH=N proton is expected to move
upfield due to the through bond effect of the hydrogen
bond, which was observed in experiment. With 1.0 equiva-
lents of Fꢀ added, the CH=N peak made a 0.2 ppm upfield
shift from d=8.51 to 8.33 ppm.
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
protons are inequivalent, as well as the two anthryl CH=N
protons. For sensors 2 and 3, similar results were observed.
The restricted conformational equilibrium was further
proved with 1H NMR spectra of sensors 1 and 12 carried
out at 508C, in which the two inequivalent peaks of the
CH=N group collapsed into one. For sensor 11, however,
there is no such restricted rotation effect in the 1H NMR
After 5 equivalents of Fꢀ was added, the proton peak of
FHFꢀ was observed at d=16.10 ppm (Figure 11g), which
supports the above proposed deprotonation mechanism.[24]
2888
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2880 –2892

Hydrazones as Fluoride Chemosensors
FULL PAPER
It should be pointed out that this does not necessarily mean
that the deprotonation only occurred at such a high concen-
tration of Fꢀ, rather, that the deprotonation is highly likely
to occur at much lower Fꢀ concentration (the UV/Vis ab-
sorption spectral changes support this assumption) and only
at high Fꢀ concentration, is the FHFꢀ complex sufficiently
kinetically stable and its abundance high enough to be de-
tected by 1H NMR.
sion). The high acidity of the CH=N proton is possibly due
to the unique molecular structure of sensor 9 and the syner-
gistic electron withdrawing effect of the two nitro groups ap-
pended at the end of the molecular framework, an effect
which is missing in the other sensors described in this paper.
Another indirect but solid piece of evidence for the
CH···F interaction comes from sensor 10, for which the titra-
tion with Fꢀ gave no UV/Vis absorption spectral changes at
all, due to the lack of the essential CH=N protons.
The 1H NMR titration of compound 1 with up to ten
ꢀ
equivalents of HSO4 failed to give any changes (data not
shown here). This means the result illustrated in Figure 11 is
specific to Fꢀ. This finding is consistent with the UV/Vis ab-
sorption results.
Switch-on fluorescent chemosensor for Fꢀ and Acꢀ: Con-
cerning the sensitivity of chemosensing, a switch-on, rather
than a switch-off fluorescent sensor is preferred. Unfortu-
nately, except for a few cases of switch-on fluorescent Fꢀ
sensors,[39,43] most others, especially the urea- and thiourea-
based anion chemosensors, are switch-off fluorescent che-
mosensors, or non-fluorescent sensors, due to the heavy
atom effect of the sulfur atom, or the binding-induced, pho-
toinduced electron-transfer (PET) quench mechanism.[20,25]
Herein we show that with the thiourea as the binding
moiety, a switch-on fluorescent chemosensor for anions was
constructed by a simple approach.[44]
The NMR titration of 12 with Fꢀ was also studied and the
result is similar to that of 1, offering further proof that the
ꢀ
ꢀ
ACHTREUNG CH=N hydrogen is able to form a CH···F hydrogen bond.
As no FHFꢀ species was found for sensor 12 even with ten
equivalents Fꢀ, a hydrogen bond or poor deprotonation
mechanism is proposed for sensor 12 in the presence of Fꢀ.
As previously discussed, a UV/Vis absorption change was
observed for sensor 9 with addition of Fꢀ (Figure 8). The
1H NMR titration of sensor 9 with Fꢀ gave additional evi-
dence for a strong CH···F interaction (Figure 12). The result
For sensor 12, which contains the fluorophore of anthra-
cene, fluorescent enhancement was observed in the presence
of Fꢀ (Figure 13). The binding constant was determined as
Figure 13. Fluorescent emission changes of sensor 12 (1.0”10ꢀ5
m in
MeCN) upon addition of Bu4NF (excited at the isosbestic point of
365 nm of the UV/Vis absorption spectra, excitation slit and emission slit
width is 5 nm), at 208C. The inset shows the best fit of the experimental
data of the fluorescent intensity at 442 nm versus equivalents of Fꢀ
added.
Figure 12. 1H NMR (400 MHz) spectra of sensor 9 (1 mm in [D6]DMSO)
with addition of Bu4NF.
logKX =3.52ꢁ0.05, which can be compared to the UV/Vis
titration results, which gave logKX =4.36ꢁ0.04. The differ-
ence between these two values indicates that the simple 1:1
binding models used to derive the respective values are not
strictly correct and that the UV/Vis and fluorescence
changes probe different parts of a more complicated set of
equilibria. The emission of 12 shows a 20 nm red shift, com-
pared to the emission of the anthracene moiety.[20,30] The
emission wavelength of the sensor also shows that there is
probably no complete delocalization, induced by the binding
of Fꢀ, otherwise an emission at longer wavelength would be
expected.
shows that with 0.1 equivalents of Fꢀ, no changes were ob-
served. With ten equivalents Fꢀ added, however, the singlet
peak of the CH=N protons at d=8.88 ppm disappeared and
the resonance of FHFꢀ appeared at d=16.13 ppm, providing
clear evidence that the CH=N group was deprotonated. The
large 0.21–0.35 ppm upfield shift for the ArH protons sup-
port the conclusion that a deprotonation process occurs in
the presence of Fꢀ (the larger upfield shift of 0.35 ppm for
Hb, compared to 0.21 ppm for Ha, also supports this conclu-
Chem. Eur. J. 2007, 13, 2880 –2892
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2889

J. Zhao, X. Peng et al.
A similar but lower magnitude fluorescence enhancement
was observed for 12 in the presence of Acꢀ. The binding
constant was determined as logK=3.16ꢁ0.06, which is in
general agreement with the result of the UV/Vis absorption
er, no blue color was observed, which is consistent with the
UV/Vis absorption study.
As the thiocarbonohydrazones are sensitive to Fꢀ and the
recognition system shows a water and methanol tolerance,
we are actively working along this line to develop a Fꢀ che-
mosensor that is effective in aqueous solutions.
ꢀ
method (logKX =3.79ꢁ0.05). In the presence of HSO4 and
ꢀ
H2PO4 , however, only a minor fluorescence enhancement
was observed, consistent with the UV/Vis absorption result.
To clarify that the fluorescence enhancement of sensor 12
with addition of Fꢀ was not caused by any photochemical re-
action of the anthracene, a control experiment was carried
out, during which the solution of sensor 12 was continuously
irradiated, whilst the fluorescence intensity was continuously
monitored, in the absence of Fꢀ. The result showed that the
fluorescence intensity did not change significantly. This con-
trol, together with the selective fluorescent response of 12
toward different anions, ruled out the possibility of a photo-
dimerization of the anthracene rings for sensor 12 (actually
the photodimerization of anthracene leads to fluorescence
attenuation).[8i] Thus, the fluorescence enhancement must
have been caused by the interaction of sensor 12 with Fꢀ
and Acꢀ.
Conclusion
In conclusion, a class of easily prepared yet sensitive thiocar-
bonohydrazone anion chemosensors was synthesized, and
the recognition of anions was also studied. Sensor 1 shows
good selectivity toward Fꢀ and Acꢀ. It was found that the
performance of the sensor is highly dependent on the sub-
stituent groups on the phenyl ring; an electron-withdrawing
nitro group resulted in a sensor with a high binding con-
stant. It was demonstrated that a CH···F hydrogen bond was
formed for the sensors in the presence of Fꢀ, based on the
single-crystal structures and NMR titration experiments. A
stepwise deprotonation of the thioamide framework was ob-
served for sensors 7 and 1 in the presence of Fꢀ, whereas for
the oxo amide sensor 11, only one deprotonation of the two
NH groups occurred. The sensor shows chemoselectivity for
the anions and the recognition system is tolerant to protic
solvents such as H2O and MeOH. Further work will concen-
trate on the development of Fꢀ sensors that are effective in
aqueous solution.
The fluorescent response of 2 toward Fꢀ was also studied;
a weak fluorescence of the sensor and only a quenching
effect were observed in the presence of Fꢀ.
Colorimetric study: Much attention has been paid to the de-
velopment of colorimetric chemosensors.[8k,16c,17,18,45–48] Com-
pound 1 is a sensitive colorimetric sensor for Fꢀ. With
0.4 equivalents of Fꢀ, a reddish color developed from the
colorless blank sensor solution. With 20 equivalents of Fꢀ
added, the red color turned to a vivid blue color. This step-
wise color change upon addition of Fꢀ infers a stepwise de-
protonation of the two NH groups.[16c,17] The red color of the
solution turning to blue on standing indicates that the de-
protonation is not a instantaneous process, which is in agree-
ment with the time-dependent UV/Vis absorption of sensor
1 with two equivalents of Fꢀ(data not shown here). It should
be pointed out that the color change is an acid-base neutral-
ization, not a supramolecular interaction.[16c]
Experimental Section
General procedures and materials: All reagents for syntheses were used
as received without further purification. UV/Vis spectra were recorded
on a Perkin–Elmer Lambda 35 UV/Vis Spectrophotometer and HP 8453
spectrophotometer, with a quartz cuvette (path length: 1 cm). Fluores-
cence spectra were recorded with a JASCO FP-6500 spectrometer. NMR
spectra were recorded on a Varian INOVA spectrometer (400 MHz). The
anions were used as the tetrabutylammonium salt. Spectroscopy grade
MeCN was used for the spectra measurements.
With 7, a similar color change profile was observed but a
higher Fꢀ concentration was required to achieve similar
color changes to those of 1. For example, 100 equivalents of
Fꢀ is required for sensor 7 to generate the blue color, in-
stead of 20 equivalents of Fꢀ for 1.
Titration with analytes: The UV/Vis and fluorescence spectra of the sen-
sors in the presence of analytes were recorded as increasing amounts of
analytes were added to the acetonitrile or DMSO. Titration plots were
generated by using Origin 5.0 (Microcal software). The binding constants
were calculated by using custom-written nonlinear least-squares curve-fit-
ting programs implemented within SigmaPlot 2000 (SPSS Inc.). The spe-
cies distribution diagram was constructed with Hyperquad Simulation
and Speciation Software (HySS, Protonic Software).
As shown by the previous binding constant studies, the
substitution of the sulfur atom with an oxygen atom in the
CH=X groups of the thiocarbonohydrazone, for example,
for sensors 1 and 11, will result a decreased binding constant
and only one deprotonation of the two NH groups. Accord-
ingly, the colorimetric recognition of Fꢀ with 11 produces
only a red color with up to 300 equivalents of Fꢀ.
General synthesis procedure:
A
solution of benzylaldehydes
(225.0 mmol) in ethanol (10 mL) was added to a mixture of thiocarbohy-
dride (1.06 g, 10.0 mmol) in water (50 mL). The heterogonous mixture
was refluxed with stirring for 5 h. Then the mixture was cooled to room
temperature and the solid was collected with suction. The crude product
was purified with column chromatography (silica gel, DCM/MeOH) to
give a powder material.
Sensor 1 is highly selective to Fꢀ and Acꢀ. With one
equivalent of Fꢀ and Acꢀ, a reddish color developed. Con-
versely, there were no color changes in the presence of ten
1,5-Bis(4-nitrobenzylidene) thiocarbonohydrazide (1): Yield: 55.9%; m.p.
234–2378C; yellow powder; 1H NMR (400 MHz, [D6]DMSO, 258C,
TMS): d=12.38 (s, 1H; NH), 12.03(s, 1H; NH), 8.72 (s, 1H; CH=N),
8.02–8.33 ppm (m, 9H); 13C NMR (100 MHz, [D6]DMSO, 258C, TMS):
d=175.5, 147.9, 146.7, 141.3, 140.4, 128.5, 128.1, 124.1 ppm; ESI-mass:
m/z (%): 371.0 (100) [MꢀH]ꢀ.
ꢀ
ꢀ
equivalents of Clꢀ, Brꢀ, HSO4 , and H2PO4 .
Similar studies show that the color of a solution of 12
turns from light yellow to red in the presence of Fꢀ. Howev-
2890
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2880 –2892

Hydrazones as Fluoride Chemosensors
FULL PAPER
1,5-Dibenzylidene thiocarbonohydrazide (2): Yield: 73.3%; m.p. 200–
2018C; white powder; 1H NMR (400 MHz, [D6]DMSO, 258C, TMS): d=
11.65 (brs, 2H; NH), 8.26–8.59 (brs, 2H; CH=N), 7.82 (brs, 4H; ArH),
7.42–7.49 ppm (m, 6H; ArH); 13C NMR (100 MHz, [D6]DMSO, 258C,
TMS): d=174.8, 148.6, 143.3, 134.1, 123.0, 128.7, 127.3 ppm; ESI-mass:
m/z (%): 305.1 (38) [M+Na]+, 587.2 (100) [2M+Na]+.
Acknowledgements
This work was financially supported by the National Natural Science
Foundation of China (NSFC Project 20642003), the Scientific Research
Foundation for the Returned Overseas Chinese Scholars, the Ministry of
Education and Dalian University of Technology (Project 893332 and
0204–872X32). The authors are grateful to Prof. Jingping Qu, Shang Gao
and Jun Hou for helpful discussions. We are also grateful to the anony-
mous referees for their critical and very helpful reviewing of the manu-
script.
1,5-Bis(4-N,N-dimethylaminobenzylidene)thiocarbonohydrazide
(3):
Yield: 66.3%; m.p. 245–2478C; yellow powder; 1H NMR (400 MHz,
[D6]DMSO, 258C, TMS): d=11.50 (brs, 1H; NH), 11.18 (brs, 1H; NH),
8.43 (brs, 1H; CH=N), 7.99 (brs, 1H; CH=N), 7.55–7.67 (brs, m, 4H;
ArH), 6.76 (brs, 4H; ArH), 2.98 ppm (s, 12H; NCH3); 13C NMR
(100 MHz, [D6]DMSO, 258C, TMS): d=173.8, 152.1, 149.7, 144.5, 129.6,
129.1, 1223.3, 121.9, 122.3, 121.9, 112.4, 40.4 ppm; ESI-mass: m/z (%):
369.1 (30) [M+H]+, 759.3 (100) [2M+Na]+.
[1] A. Prasanna de Silva, H. Q. Nimal Gunaratne, T. Gunnlaugsson,
A. J. M. Huxley, C. P. McCoy, J. T. Rademacher, T. E. Rice, Chem.
Rev. 1997, 97, 1515–1566.
[2] A. Robertson, S. Shinkai, Coord. Chem. Rev. 2000, 205, 157–199.
[3] P. D. Beer, P. A. Gale, Angew. Chem. 2001, 113, 502–532; Angew.
Chem. Int. Ed. 2001, 40, 486–516.
[4] M. D. Best, S. L. Tobey, E. V. Anslyn, Coord. Chem. Rev. 2003, 240,
3–15.
[5] R. Martínez-MµÇez, F. Sancenón , Chem. Rev. 2003, 103, 4419–
4476.
p-Dimethylaminobenzaldehydethiocarbonohydrazone (5): Yield: 60.8%;
m.p. 178–1808C; yellow powder; 1H NMR (400 MHz, [D6]DMSO, 258C,
TMS): d=11.19 (s, 1H; NH), 9.55 (s, 1H; NH), 7.89 (s, 1H; CH=N), 7.61
(d, J=8.0 Hz, 2H; ArH), 6.68 (d, J=8.0 Hz, 2H; ArH), 4.80 (s, 2H;
NH2), 2.96 ppm (s, 6H; CH3); ESI-mass: m/z (%): 260.0 (100) [M+Na]+,
497.2 (40) [2M+Na]+.
p-Nitrobenzaldehyde-thiocarbonohydrazone (6): Yield: 88.0%; m.p. 204–
2068C; yellow powder; 1H NMR (400 MHz, [D6]DMSO, 258C, TMS):
d=11.76 (s, 1H; NH), 10.17 (s, 1H; NH), 8.22 (d, J=8.0 Hz, 2H; ArH),
8.13 (d, J=8.0 Hz, 2H; ArH), 8.09 (s, 1H; CH=N), 4.96 ppm (s, 2H;
NH2); ESI-mass: m/z (%): 238.1 (100) [MꢀH]ꢀ, 274.0(74) [M+Cl]ꢀ.
Bisthiocarbonohydrazone of 4-nitroacetophenone (7): Yield: 82.1%; m.p.
210–2128C; yellow powder; 1H NMR (400 MHz, [D6]DMSO, 258C,
TMS): d=11.14 (brs, 2H; NH), 8.28 (d, J=8.0 Hz, 4H; ArH), 8.14 (d,
J=8.0 Hz, 4H; ArH), 2.45 ppm(s, 6H; CH3); ESI-mass: m/z (%): 399.2-
[6] L. Pu, Chem. Rev. 2004, 104, 1687–171.
[7] R. Martínez-MµÇez, F. Sancenón, J. Fluoresc. 2005, 15, 267–285.
[8] a) C. Suksai, T. Tuntulani, Top. Curr. Chem. 2005, 255, 163–198;
b) M.-L. Lehaire, R. Scopelliti, H. Piotrowski, K. Severin, Angew.
Chem. 2002,114, 1477–1480; Angew. Chem. Int. Ed. 2002, 41, 1419–
1422; c) I. H. A. Badr, M. E. Meyerhoff, J. Am. Chem. Soc. 2005,
127, 5318–5319; d) K. H. Lee, H. Y. Lee, D. H. Lee, J. I. Hong, Tet-
rahedron Lett. 2001, 42, 544–5449; e) R. Miao, Q. Y. Zheng, C. F.
Chen, Z. T. Huang, Tetrahedron Lett. 2004, 45, 4959–4962; f) Z. Q.
Liu, M. Shi, F. Y. Li, Q. Fang, Z. H. Chen, T. Yi, C. H. Huang, Org.
Lett. 2005, 7, 5481–5484; g) M. Melaimi, F. P. Gabbai, J. Am. Chem.
Soc. 2005, 127, 9680–9681; h) J. S. Wu, J. H. Zhou, ; P. F. Wang,
X. H. Zhang, S. K. Wu, Org. Lett. 2005, 7, 2133–2136; i) K. J. Wal-
lace, W. J. Belcher, D. R. Turner, K. F. Syed, J. W. Steed, J. Am.
Chem. Soc. 2003, 125, 9699–9715; j) T. Neumann, Y. Dienes, T.
Baumgartner, Org. Lett. 2006, 8, 495–497; k) H. Maeda, Y. Kusu-
nose, Chem. Eur. J. 2005, 11, 5661–5666; l) T. Gunnlaugsson, P. E.
Kruger, P. Jensen, J. Tierney, H. Dato Paduka Ali, G. M. Hussey, J.
Org. Chem. 2005, 70, 10875–10878.
[9] M. Shionoya, H. Furuta, V. Lynch, A. Hamiman, J. L. Sessler, J.
Am. Chem. Soc. 1992, 114, 5714–5722.
[10] C. B. Black, B. Andrioletti, A. C. Try, C. Ruiperez, J. L. Sessler, J.
Am. Chem. Soc. 1999, 121, 10438–10439.
[11] P. Anzenbacher, Jr., K. Jursikova, J. L. Sessler, J. Am. Chem. Soc.
2000, 122, 9350–9351.
[12] a) M. Takeuchi, T. Shioya, T. M. Swager, Angew. Chem. 2001, 113,
3476–3480; Angew. Chem. Int. Ed. 2001, 40, 3372–3376; b) C.
Bohne, H. Ihmels, M. Waidelich, C. Yihwa, J. Am. Chem. Soc. 2005,
127, 17158–17159; c) T. Mizuno, W.-H. Wei, L. R. Eller, J. L. Sess-
ler, J. Am. Chem. Soc. 2002, 124, 1134–1135.
[13] a) Y. Kubo, M. Yamamoto, M. Ikeda, M. Takeuchi, S. Shinkai, S. Ya-
maguchi, K. Tamao, Angew. Chem. 2003, 115, 2082–2086; Angew.
Chem. Int. Ed. 2003, 42, 2036–2040; b) S. Arimori, M. G. Davidson,
T. M. Fyles, T. G. Hibbert, T. D. James, G. I. Kociok-Kçhn, Chem.
Commun. 2004, 1640–1641; c) N. Di Cesare, J. R. Lakowicz, Anal.
Biochem. 2002, 301, 111–116.
A
Bis(4-nitrobenzylidene)hydrazine (9): Yield: 63.7%; m.p. >3008C; light
1
yellow powder; H NMR (400 MHz, [D6]DMSO, 258C, TMS): d=8.88 (s,
2H; CH=N), 8.37 (d, J=8.0 Hz, 4H; ArH), 8.16 ppm (d, J=8.0 Hz, 4H;
ArH); ESI-mass: m/z (%): 298.3 (100) [M+H]+.
Bis-[1-(4-nitro-phenyl)ethylidene]hydrazine (10): Yield: 8.8%; m.p.
189.3–190.78C; orange powder; 1H NMR (400 MHz, [D6]DMSO, 258C,
TMS): d=8.31 (d, J=8.0 Hz, 4H; ArH), 8.17 (d, J=8.0 Hz, 4H; ArH),
2.33 ppm (s, 6H; CH3); ESI-mass: m/z (%): 326.3 (100) [MꢀH]ꢀ.
1,5-Bis(4-nitrobenzylidene)carbonohydrazide (11): Yield: 43.5%; m.p.
2528C; light yellow powder; 1H NMR (400 MHz, [D6]DMSO, 258C,
TMS): d=11.22 (brs, 2H; NH), 8.28 (m, 6H), 8.02 ppm (d, J=12.0 Hz,
4H; ArH); ESI-mass: m/z (%): 356.5 (100) [M+H]+.
1,5-Bis(9-anthracylidine) thiocarbonohydrazide (anthracene-9-carbalde-
hydebisthiocarbohydrazone) (12): Yield: 87.0%; m.p. 212–2148C; orange
powder; 1H NMR (400 MHz, [D6]DMSO, 258C, TMS): d=12.12–12.26
(brs, 2H; NH), 9.71 (brs, 1H; CH=N), 9.45 (brs, 1H; CH=N), 8.60–
8.97 (m, 6H; ArH), 8.16–8.18 (m, 4H; ArH), 7.59–7.65 ppm(m, 8H;
ArH); ESI-mass: m/z (%): 482.9 (100) [M+H]+.
X-ray crystallography: The single-crystal structure solution was per-
formed by using SHELXL-97.[49] The single crystal of sensor 2 was grown
by keeping a concentrated solution of the sensor in DMSO in a refrigera-
tor for three months. Crystal data for 2: light yellow column, C15H14N4S,
Mr =282.4, tetragonal, space group Pbca, a=8.291(3), b=18.644(6), c=
18.644 Š, a=90.00, b=90.00, g=90.008, V=2881.94 Š3, reflections col-
lected: 14899; independent reflections: 2949 (Rint =0.0399); final R indi-
ces [I>s(I)]: R1 =0.0399, wR2 =0.0994; R indices (all data): R1 =0.0583,
wR2 =0.1098. The single crystal of sensor 12 was grown by slow evapora-
tion of a dilute solution of the sample in THF. Crystal data for 12: light
[14] J. Y. Kwon, Y. J. Jang, S. K. Kim, K. H. Lee, J. S. Kim, J. Yoon, J.
Org. Chem. 2004, 69, 5155–5157.
¯
yellow prism, C31H22N4S, Mr =482.6, triclinic, space group P1, a=
9.027(2), b=10.722(3), c=15.894(4) Š, a=78.664(3), b=74.134(4), g=
80.434(4)8, V=1440.51 Š3, reflections collected: 7845; independent re-
flections: 5550 (Rint =0.0149); final R indices [I>s(I)]: R1 =0.0490, wR2 =
0.1217; R indices (all data): R1 =0.0743, wR2 =0.1340. CCDC-612122 (2)
and CCDC-607150 (12) contain the supplementary crystallographic data
for this paper. These data can be obtained free of charge from The Cam-
bridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif.
[15] T.-H. Kim, T. M. Swager, Angew. Chem. 2003, 115, 4951–4954;
Angew. Chem. Int. Ed. 2003, 42, 4803–4806.
[16] a) M. Vµzquez, L. Fabbrizzi, A. Taglietti, R. M. Pedrido, A. M. Gon-
zµlez-Noya, M. R. Bermejo, Angew. Chem. 2004, 116, 1996–1999;
Angew. Chem. Int. Ed. 2004, 43, 1962–1965; b) Y.-M. Zhang, W.-X.
Xu, Y.-Q. Zhou, H. Yao, T.-B. Wei, Acta. Chimica. Sinica. 2006, 64,
79–84; c) V. Amendola, D. Esteban-Gomez, L. Fabbrizzi, M. Lic-
chelli, Acc. Chem. Res. 2006,39, 343–353.
Chem. Eur. J. 2007, 13, 2880 –2892
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2891

J. Zhao, X. Peng et al.
[17] D. Esteban-Gomez, L. Fabbrizzi, M. Licchelli, J. Org. Chem. 2005,
70, 5717–5720.
[18] M. Boiocchi, L. D. Boca, D. E. Gómez, L. Fabbrizzi, M. Licchelli, E.
Monzani, J. Am. Chem. Soc. 2004, 126, 16507–16514.
[19] E. J. Cho, B. J. Ryu, Y. J. Lee, K. C. Nam, Org. Lett. 2005, 7, 2607–
2609.
[34] a) M. P. M. Marques, A. M. Amorim da Costa, P. J. A. Ribeiro-
Claro, J. Phys. Chem. A 2001, 105, 5292–5297; b) Y. Gu, T. Kar, S.
Scheiner, J. Am. Chem. Soc. 1999, 121, 9411–9422.
[35] V. R. Thalladi, H.-C. Weiss, D. Blꢂser, R. Boese, A. Nangia, G. R.
Desiraju, J. Am. Chem. Soc. 1998, 120, 8702–8710.
[36] K. N. Houk, S. Menzer, S. P. Newton, F. M. Raymo, J. F. Stoddart,
D. J. Williams, J. Am. Chem. Soc. 1999, 121, 1479–1487.
[37] C. Schiene-Fischer, G. Fischer, J. Am. Chem. Soc. 2001, 123, 6227–
6231.
[38] C. F. Chen, Q. Y. Chen, Tetrahedron Lett. 2004, 45, 3957–3960.
[39] a) D. H. Lee, J. H. Im, J. H. Lee, J. I. Hong, Tetrahedron Lett. 2002,
43, 9637–9640; b) J. Y. Lee, E. J. Cho, S. Mukamel, K. C. Nam, J.
Org. Chem. 2004, 69, 943–950.
[40] D. Jimꢃnez, R. Martínez-MµÇez, F. Sancenón, J. Soto, Tetrahedron
Lett. 2002, 43, 2823–2825.
[41] D.-W. Yoon, H. Hwang, Ch. H. Lee, Angew. Chem. 2002, 114, 1835–
1837; Angew. Chem. Int. Ed. 2002, 41, 1757–1759.
[42] S. Ghosh, A. R. Choudhury, T. N. Guru Row, U. Maitra, Org. Lett.
2005, 7, 1441–1444.
[43] a) V. Thiagarajan, P. Ramamurthy, D. Thirumalai, V. T. Ramakrishn-
an, Org. Lett. 2005, 7, 657–660; b) S. K. Kim, J. H. Bok, R. A.
Bartsch, J. Y. Lee, J. S. Kim, Org. Lett. 2005, 7, 4839–4842; c) H.
Tong, L. Wang, X. Jing, F. Wang, Macromolecules 2003, 36, 2584–
2586.
[44] G. Xu, M. A. Tarr, Chem. Commun. 2004, 1050–1051.
[45] S. Yamaguchi, S. Akiyama, K. Tamao, J. Am. Chem. Soc. 2001, 123,
11372–11375.
[46] D. Saravanakumar, N. Sengottuvelan, M. Kandaswamy, P. G. Ara-
vindanb, D. Velmurugan, Tetrahedron Lett. 2005, 46, 7255–7258.
[47] a) P. Anzenbacher, Jr., M. A. Palacios, K. Jursikova, M. Marquez,
Org. Lett. 2005, 7, 5027–5030; b) X. He, S. Hu, K. Liu, Y. Guo, J.
Xu, S. Shao, Org. Lett. 2006, 8, 333–336.
[48] a) D. A. Jose, D. K. Kumar, B. Ganguly, A. Das, Org. Lett. 2004, 6,
3445–3448; b) T. Ghosh, B. G. Maiya, M. W. Wong, J. Phys. Chem.
A 2004, 108, 11249–11259.
[20] T. Gunnlaugsson, A. P. Davisab, M. Glynna, Chem. Commun. 2001,
2556–2557.
[21] a) T. Sifferlen, M. Rueping, K. Gademann, B. Jaun, D. Seebach,
Helv. Chim. Acta 1999, 82, 2067–2093; b) R. Frank, M. Jakob, F.
Thunecke, G. Fischer, M. Schutkowski, Angew. Chem. 2000, 112,
1163–1165; Angew. Chem. Int. Ed. 2000, 39, 1120–1122;
[22] a) L. Nie, Z. Li, J. Han, X. Zhang, R. Yang, W. X. Liu, F. Y. Wu,
J. W. Xie, Y. F. Zhao, Y. B. Jiang, J. Org. Chem. 2004, 69, 6449–
6454; b) A. Bacchi, A. Bonini, M. Carcelli, F. Ferraro, E. Leporati,
C. Pelizzi, G. Pelizzi, J. Chem. Soc. Dalton Trans. 1996, 2699–2704.
[23] a) J. Zhao, J.-C. Micheau, C. Vargas, C. Schiene-Fischer, Chem.Eur.
J. 2004, 23, 6093–6101; b) G. Scherer, M. L. Kramer, M. Schutkow-
ski, U. Reimer, G. Fischer, J. Am. Chem. Soc. 1998, 120, 5568–5574.
[24] X. Peng, Y. Wu, J. Fan, M. Tian, K. Han, J. Org. Chem. 2005, 70,
10524–10531.
[25] T. Gunnlaugsson, P. E. Kruger, T. C. Lee, R. Parkesh, F. M. Pfeffera,
G. M. Husseya, Tetrahedron Lett. 2003, 44, 6575–6578.
[26] T. J. Ryan, G. Lecollinet, T. Velasco, A. P. Davis, Proc. Natl. Acad.
Sci. USA 2002, 99, 4863–4866.
[27] C. H. Lee, H. K. Na, D. W. Yoon, D. H. Won, W. S. Cho, V. M.
Lynch, S. V. Shevchuk, J. L. Sessler, J. Am. Chem. Soc. 2003, 125,
7301–7306.
[28] K. Sato, S. Arai, T. Yamagishi, Tetrahedron Lett. 1999, 40, 5219–
5222.
[29] S. Yun, H. Ihm, H. G. Kim, C.-W. Lee, B. Indrajit, K. S. Oh, Y. J.
Gong, J. W. Lee, J. Yoon, H. C. Lee, K. S. Kim, J. Org. Chem.
2003,68, 2467–2470.
[30] J. Zhao, M. G. Davidson, M. F. Mahon, G. Kociok-Kohn, T. D.
James, J. Am. Chem. Soc. 2004, 126, 16179–16186.
[31] A. Zhang, L. Zhang, Y. Hou, Y. Xiong, X. Li, K. Yu, Youji Huaxue
(Chin. J. Org. Chem.) 2003, 23, 281–285.
[32] H. Benmansour, R. D. Chambers, G. Sandford, S. Dahaoui, D. S.
Yufita, Arkivoc. 2000, 1, 202–209.
[49] G. M. Sheldrick, SHELXL-97, Computer Program for Crystal Struc-
ture Refinement, University of Gçttingen; Gçttingen (Germany),
1997
[33] a) R. D. Gautam, Acc. Chem. Res. 1991, 24, 290–296; b) J. W. Steed,
J. L. Atwood, Supramolecular Chemistry, Wiley, New York, 2000,
pp. 434–437; c) S. Tomꢁs, R. Prohens, M. Vega, M. Rotger, M. Deyꢁ
Pere, P. Ballester, A. Costa, J. Org. Chem. 1996, 61, 9394–9401.
Received: June 25, 2006
Revised: October 25, 2006
Published online: January 2, 2007
2892
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2880 –2892