Ferrocenyl compound as a multiresponsive calcium chemosensor with remarkable fluorescence properties in CH3CN

Article abstract of DOI:10.1021/ic060535e

We have synthesized a novel disubslituted ferrocenyl compound [Fe(C ₅H₄CO(CH=CH)₂C₆H₄NEt ₂)₂] (3) that displays a remarkable fluorescence quantum yield (1.1 × 10^-1) in acetonitrile, and we have studied its capacity for calcium detection in depth using both electrochemical and optical techniques in this medium. The results of our NMR analysis reveal that the ligand-calcium interaction is CO-centered and that an uncommon equilibrium occurs between 3 and calcium triflate, involving five species of different stoichiometries. In contrast, our analysis of the UV-vis absorption data indicates that only three species of different stoichiometries are formed when calcium perchlorate is used with 3. Mass spectrometry measurements provide strong support for the formation of all these different species in solution. In addition, the electrochemical detection of calcium triflate by 3 leads to an irreversible Fe^II/ Fe^III oxidation process with an unusual negative shift (-60 mV) caused by the ⁿBu₄NBF₄ salt effect on the Ca²⁺-3 interaction process. Compound 3 can also be an original optical probe to detect calcium perchlorate over a wide range of salt concentration by UV-vis absorption spectroscopy. The most original and intriguing property of compound 3 is that it exhibits an unprecedented multistep fluorescence behavior upon addition of this salt.

Full text of DOI:10.1021/ic060535e

Inorg. Chem. 2006, 45, 5691−5702
Ferrocenyl Compound as a Multiresponsive Calcium Chemosensor with
Remarkable Fluorescence Properties in CH3CN
Be´atrice Delavaux-Nicot,*,† Je´roˆme Maynadie´,† Dominique Lavabre,‡ and Suzanne Fery-Forgues‡
Laboratoire de Chimie de Coordination du CNRS, UPR 8241, 205, route de Narbonne, F-31077,
Toulouse Cedex 04, France, and Laboratoire des Interactions Mole´culaires et Re´actiVite´ Chimique
et Photochimique, UMR 5623 du CNRS, UniVersite´ Paul Sabatier, 118 route de Narbonne,
F-31062, Toulouse Cedex, France
Received March 29, 2006
We have synthesized a novel disubstituted ferrocenyl compound [Fe(C5H4CO(CH
a remarkable fluorescence quantum yield (1.1
10-1) in acetonitrile, and we have studied its capacity for calcium
detection in depth using both electrochemical and optical techniques in this medium. The results of our NMR
analysis reveal that the ligand calcium interaction is CO-centered and that an uncommon equilibrium occurs between
3 and calcium triflate, involving five species of different stoichiometries. In contrast, our analysis of the UV vis
dCH)2C6H4NEt2)2] (3) that displays
×
−
−
absorption data indicates that only three species of different stoichiometries are formed when calcium perchlorate
is used with 3. Mass spectrometry measurements provide strong support for the formation of all these different
species in solution. In addition, the electrochemical detection of calcium triflate by 3 leads to an irreversible FeII/
n
+
FeIII oxidation process with an unusual negative shift (
−
60 mV) caused by the Bu4NBF4 salt effect on the Ca2
−3
interaction process. Compound 3 can also be an original optical probe to detect calcium perchlorate over a wide
range of salt concentration by UV vis absorption spectroscopy. The most original and intriguing property of compound
3 is that it exhibits an unprecedented “multistep” fluorescence behavior upon addition of this salt.
−
Introduction
sensors is to synthesize a molecule that is capable of reporting
on the recognition event through a variety of physical
responses. This allows thorough information to be obtained
by comparing the different physical responses gathered upon
addition of an analyte. It also allows the same sensor to be
used in various experimental conditions and, possibly, for
the detection of different analytes.2 In this context, the
coupling between electrochemical and optical techniques, in
particular UV-vis absorption spectroscopy, has recently
attracted much attention.1b,3 An exciting advantage of these
systems could be the tuning of the optical properties via
oxidation state modulation of the redox center.4 In this
research area, ferrocenyl derivatives are of special interest:
(i) from a synthetic standpoint, ferrocene is a very convenient
building block for a redox active ligand because it can
relatively easily be functionalized and incorporated by many
Sensing of ions is an area of increasing research activity1
because ions play a fundamental role in biology, chemical
processes, and environmental pollution. Sensing systems are
usually composed of a signaling unit linked to a receptor,
so that ion binding is determined by a measurable physical
change. An attractive way of achieving a new generation of
* To whom correspondence should be addressed. E-mail:
delavaux@lcc-toulouse.fr. Fax: 33 (0)5 61 55 30 03.
† Laboratoire de Chimie de Coordination du CNRS.
‡ Universite´ Paul Sabatier.
(1) For some examples, see: (a) Beer, P. D. AdV. Inorg. Chem. 1992 39,
79-157. (b) Beer, P. D.; Cadman, J. Coord. Chem. ReV. 2000, 205,
131-155. (c) Delavaux-Nicot, B.; Bigeard, A.; Bousseksou, A.;
Donnadieu, B.; Commenges, G. Inorg. Chem. 1997, 36, 4789-4797
and references therein. (d) Astruc, D.; Daniel, M. C.; Ruiz, J. Chem.
Commun. 2004, No. 23, 2637-2649. (e) Ion, A. I.; Ion, J.-C.; Pailleret,
A.; Popescu, A.; Saint-Aman, E.; Ungureanu, E.; Siebert, E.; Ziessel,
R. Sens. Actuators 1999, B59, 118-212. (f) Miyaji, H.; Collinson, S.
R.; Prokes, I.; Tucker, J. H. R. Chem. Commun. 2003, No. 1, 64-65.
(g) de Silva, A. P.; Gunaratne, H. Q. N.; Gunnlaugsson, T.; Huxley,
A. J. M.; McCoy, C. P.; Rademacher, J. T.; Rice, T. E. Chem. ReV.
1997, 97, 1515-1566. (h) Czarnik, A. Fluorescent Chemosensors for
Ion and Molecule Recognition; ACS Symposium Series 538; American
Chemical Society: Washington, DC, 1993. (i) Tsien, R. Y. In Methods
in Cell Biology; Taylor, D., Wang, Y.-L, Eds.; Academic Press:
London, 1989; Vol. 30, p 127.
(2) Jimenez, D.; Martinez-Manez, R.; Sancenon, F.; Ros-Lis, J. V.; Soto,
J.; Benito, A.; Garcia-Breijo, E. Eur. J. Inorg. Chem. 2005, 2393-
2403.
(3) (a) Beer, P. D.; Hayes, E. J. Coord. Chem. ReV. 2003, 240, 167-189.
(b) Lopez, J. L.; Tarraga, A.; Espinosa, A.; Velasco, M. D.; Molina,
P.; Lloveras, V.; Vidal-Gancedo, J.; Rovira, C.; Veciana, J.; Evans,
D. J.; Wurst, K. Chem.sEur. J. 2004, 10, 1815-1826.
10.1021/ic060535e CCC: $33.50
Published on Web 06/06/2006
© 2006 American Chemical Society
Inorganic Chemistry, Vol. 45, No. 14, 2006 5691

Delavaux-Nicot et al.
Scheme 1. Compounds 1 and 2
structures, and (ii) its electrochemical and UV-vis spectro-
scopic properties can be perturbed by the proximity of bound
guests. We are especially interested in the fluorescence
properties of this kind of derivative. Actually, fluorimetry
is presently the optical method that offers the highest
sensitivity, and ion recognition by fluorescent sensors has
demonstrated the usefulness of this technique. However,
ferrocenyl receptors which contain both electroactive and
fluorescent signaling units are still rare,5 probably, because
ferrocene derivatives are known to be efficient fluorescence
quenchers. 5c,6 Even so, this kind of compound could be the
keystone of new families of ion chemosensors, and it has
already been announced that the fabrication of these systems
and their integration into different supports would certainly
lead to novel prototype molecular sensory devices for
commercial applications.1b
Although the field of research dealing with fluorescent
and electrochemical ferrocenyl ion sensors is still in its
infancy, a few interesting ligands have already been
designed.5a,7,8,9,10 Let us remark that all the “purely” ferro-
cenyl receptors8-10 consist of well-separated three component-
systems. They contain one or several ferrocene units, a
complexing moiety, and one or several fluorescent moieties,
which are connected together by saturated bonds. In the
absence of guest, the ferrocene residue acts as a quencher
for the fluorescent unit, through the involvement of a
photoinduced electron-transfer process. In the presence of
an anion, the fluorescence revival observed is the result of
the inhibition of this mechanism.
As early as 1999, we first reported a totally different type
of ligand, where the ferrocenyl, complexing, and fluorescent
moieties belong to the same conjugated π-electron system.
To do so, we have investigated the synthesis of electroactive
receptors, which combine a ferrocenyl unit and a purely
organic fluorescent ion sensor subunit containing an R-amino
complexing moiety (-COCHdCHC6H4-pR).5b Of particular
interest is the fact that the monosubstituted compound
[(C5H5)Fe(C5H4COCHdCHC6H4NEt2)] (2) is not fluorescent,
while the disubstituted compound [Fe(C5H4COCHdCHC6H4-
NEt2)2] (1) (Scheme 1) displays remarkable fluorescence
properties and behaves as a new type of optically responsive
calcium-sensing device in CH3CN.11 In this compound, the
ferrocenyl moiety is far from being a simple fluorescence
quencher and acts essentially as an auxochrome.
Since then, other groups have synthesized new conjugated
ferrocenyl systems with interesting luminescence properties.12
A remarkable example is that of Miranda and Molina’s team
who has recently reported that the addition of 1 equiv of
Li+ to a ferrocene-anthracene-linked dyad remarkably
increases its fluorescence intensity (7.3-fold) in an acetoni-
trile/water (70:30) mixture at pH 5.12a
In this promising research field, any accurate or direct
comparison concerning the ion detection performances of
the few ferrocenyl existing systems is still difficult as
numerous parameters involved in the “multisensing phen-
omenon”, such as the medium, the technique considered, or
the nature and quantity of the ion, may vary and have to be
considered. This is the reason why the introduction focused
only on some structural differences that could, in particular,
be responsible of the different natures of optical sensing.
In the present study, we are reporting the design of a new
ferrocenyl derivative, in which we have retained the disub-
stituted topology of compound 1 but increased the length of
the unsaturated chain. Our goal was 2-fold: first, we wished
to obtain another unusual cation sensor with interesting
fluorescence and electrochemical properties, and second, we
expected to gain a better understanding of the factors that
could influence the physical properties in this family.
Moreover, we were especially interested in the processes
leading to the detection phenomena. Therefore, we have
successfully synthesized compound 3 (Scheme 2) and studied
two complementary important aspects of its chemistry, which
included the following: (1) the capacity of 3 in detecting
calcium both by electrochemical and optical techniques in
CH3CN and (2) the type of interaction involved with this
cation. So, in the following sections, we are describing the
preparation and characterization of ligand 3, and the study
of its physicochemical properties without and with calcium
salt. In addition to electrochemical and fluorescence mea-
(4) (a) Harvey, P. D.; Gan, L.; Aubry, C. Can J. Chem. 1990, 68, 2278-
2288. (b) Harvey, P. D.; Gan, L. Inorg. Chem. 1991, 30, 3239-3241.
(c) Cuffe, L.; Hudson, R. D. A.; Gallagher, J. F.; Jennings, S.;
McAdam, C. J.; Connelly, R. B. T.; Manning, A. R.; Robinson, B.
H.; Simpson, J. Organometallics 2005, 24, 2051-2060. (d) McGale,
E.M.; Robinson, B. H.; Simpson, J.; Organometallics 2003, 22, 931-
939. (e) McAdam, C. J.; Morgan, J. L.; Robinson, B. H.; Simpson, J.
Organometallics 2003, 22, 5126-5136.
(5) (a) Beer, P. D.; Szemes, F.; Balzani, V.; Sala`, C. M.; Drew, M. G. B.;
Dent, S. W.; Maestri, M. J. Am. Chem. Soc. 1997, 119, 11864-11875.
(b) Delavaux-Nicot, B.; Fery-Forgues, S. Eur. J. Inorg. Chem. 1999,
1821-1825. (c) Fery-Forgues, S.; Delavaux-Nicot, B. J. Photochem.
Photobiol. A 2000, 132, 137-159.
(6) Geoffroy, G. L.; Wrighton, M. S. Organometallic Photochemistry;
Academic Press: New York, 1979; Chapter 5, pp 230-257.
(7) Beer, P. D.; Graydon, A. R.; Sutton, L. R. Polyhedron 1996, 15, 2457-
2461.
(8) Sancenon, F.; Benito, A.; Hernandez, F. J.; Lloris, J. M.; Martinez-
Manez, R.; Pardo, T.; Soto, J. Eur. J. Inorg. Chem. 2002, 866-875.
(9) Oton, F.; Tarraga, A.; Velasco, M. D.; Espinosa, A.; Molina, P. Chem.
Commun. 2004, 1658-1659.
(10) Kuo, L.-J.; Liao, J.-H.; Chen, C.-T.; Huang, C.-H.; Chen, C.-H.; Fang,
J.-M. Org. Lett. 2003, 5, 1821-1824.
1
surements, we have also performed thorough H NMR and
UV-vis investigations concerning the interaction of com-
(11) Maynadie´, J.; Delavaux-Nicot, B.; Fery-Forgues, S.; Lavabre, D.;
Mathieu, R. Inorg. Chem. 2002, 41, 5002-5004.
(12) For examples, see: (a) Caballero, A.; Tormos, R.; Espinosa, A.;
Velasco, M. D.; Tarraga, A.; Miranda, M. A.; Molina, P. Organic
Lett. 2004, 4599-4602. (b) Butler, I. R.; Callabero, A. G.; Kelly, G.
A.; Amey, J. R.; Kraemer, T.; Thomas, D. A.; Ligth, M. E.; Gelbrich,
T.; Coles, S. J. Tetrahedron Lett. 2004, 467-472.
5692 Inorganic Chemistry, Vol. 45, No. 14, 2006

Chemosensor with Remarkable Fluorescence
Scheme 2. Synthetic Route for Compounds 3 and 4a
a Reagents and conditions: (a) NaOH 1 equiv, ethanol, 20 °C; (b) HBF4‚Et2O 2 equiv, CH3CN, 20 °C.
Table 1. 1H NMR and 13C Shift Variations (∆(δ), ppm) of Selected
Groups of Compound 3 upon (a) Protonation and (b) Calcium Addition
(2 equiv)a
CHa CHb CHc CHd CHp CHe CHf CHg CHh CO
(a) 3 + 2H+
1H
0.19
3.70 -2.33
0.00
0.36
0.09
0.93
9.20
0.26 0.05 0.07
7.62 0.26 0.81
0.33
0.09
13C
7.66 -4.28 0.04
(b) 3 +2Ca2+
0.28 -0.14 -0.09 0.00 0.16 0.21 -0.10
6.11 0.81 -0.21 -0.07 1.26 1.44 -0.67
Figure 1. 1H NMR (400 MHz, CD3CN) spectra of 3 in the 4-8 ppm
range.
1H -0.12
13C -2.11
0.06
4.49 2.24
a [L] ) 5 ×10-3 M; see group labeling in Scheme 2; 1H NMR {400.13
MHz} in CD3CN at 293K; (-) upfield shift.
pound 3 with calcium to quantify the interaction process that
occurs, and we discuss our results. We also present the
preparation and characterization of the protonated derivative
[3H2][BF4]2 (4) which is a key compound in the calcium
triflate electrochemical detection process. When useful to
make our point, we have compared our present series of data
with those previously reported for compounds 1 and 2 and
with those of some derivatives of 2. Finally, in the last
section, we highlight the remarkable fluorescence behavior
of compound 3 upon calcium addition.
Upon treatment of a CH3CN solution of compound 3 with
HBF4‚Et2O in a 1:2 stoichiometry, the solution turned from
red to pink, affording the protonated species [3H2][BF4]2 (4).
Compound 4 was isolated in good yield (68%). Its charac-
1
terization was fully achieved by H and 13C 2D NMR
experiments. In the 1H NMR spectrum (CD3CN, 400 MHz),
the protonation reaction was confirmed by the appearance
of a new signal at 8.95 ppm attributed to the proton of the
NH+ group.1c,14 The NCH3 groups appeared at a downfield-
shifted position lower than those of 3 (0.26 ppm), which is
consistent with the quaternarization of the nitrogen atom. In
Results and Discussion
Synthesis and Characterization of Compound 3 and
of its Protonated Derivative 4. The synthesis of compound
3 was realized by reaction of the 1,1′-diacetylferrocene with
2 equiv of the appropriate aldehyde in basic medium (Scheme
2). The product was isolated in a good yield (65%) as a red-
orange powder. Its structure was deduced from spectroscopic
NMR data, which also revealed the perfect symmetry of the
molecule with the protons of the two branches giving exactly
the same signal. The 1H NMR spectra of 3 in the 4-8 ppm
range is provided in Figure 1. 2D NMR experiments at 400
MHz were undertaken in CD3CN to provide a complete
assignment of each signal. The IR spectrum exhibited a ν-
(CdO) vibration at 1645 cm-1, this low value being the result
of the conjugation of this group with the rest of the
molecule.13 The elemental analyses and mass spectra were
in agreement with the formula proposed for 3 in Scheme 2
(see Experimental Section).
1
Table 1a, the variations of the H and 13C NMR shift show
that the molecule is totally affected by the protonation of
the nitrogen atom.
NMR Study of the 3-Ca2+ Interaction. A. Nature of
the Interacting Sites. Since we will show in the next section
that compound 3 detects calcium salt electrochemically, we
have chosen to examine the 3-Ca2+ interaction process
responsible for this detection. First, to determine the nature
1
of the interacting site, we recorded the H NMR spectra of
compound 3 in CD3CN in the presence of 2 equiv of calcium
salt. Next, to properly ascertain the variations of the shift
observed, we have performed 2D NMR experiments.
As we indicated in Table 1b,1H NMR shift variations were
observed along the whole conjugated link, with the strongest
shift being associated with the Hb protons. In contrast to
the protonation reaction, calcium addition did not induce
drastic shifts in the vicinity of the nitrogen atom, and the
Hp protons were unaffected. The 13C NMR shift variations
(13) Silverstein, R. M.; Bassler, G. C.; Morril, T. C. In Spectrometric
Identification of Organic Compounds, 4th ed.; J. Wiley and Sons: New
York, 1981; pp 117-118.
Inorganic Chemistry, Vol. 45, No. 14, 2006 5693

Delavaux-Nicot et al.
Table 2. Association Constants Related to the Different Formed
Species with Ligand 3 and Calcium Triflate in Acetonitrile, Determined
by Processing the NMR Dataa
3M
(3)2M
3M2
(3)3M
3M3
K1 (M-1
)
K2 (M-1
)
K3 (M-1
)
K4 (M-1
)
K5 (M-1
)
2.07 × 102
6.41 × 103
4.37 × 102
1.11 × 103
4.93 × 101
a Values given with (15% error.
molecule including the Cp rings contributes to the electronic
interaction with the cation.
The curves (Figure 2) were then processed according to a
global curve-fitting method previously reported,17,18,19 and
the analysis was performed simultaneously for the six protons
considered. Complete agreement between theory and experi-
ment could only be obtained by taking into account the
existence of five species with different stoichiometries. They
are formed by compound 3 with calcium and can be written
as: 3M, (3)2M, 3M2, (3)3M, and 3M3 (where M is the
calcium cation). Their corresponding association constants
(Kn) are given in Table 2. In Figure 2 (top), we displayed
1
the H NMR experimental data points together with the
calculated curves to show the quality of the fit. More
precisely, it must be noted that no correct fit (see S2) could
be obtained without considering the (3)2M and (3)3M species.
Then, taking into account the 3M2 and 3M3 species allowed
the residual error, Er (see Experimental Section), to be
significantly reduced from 2.9 × 10-5, with 3 species, to 4
× 10-6, with 5 species.
As shown in Figure 2 (bottom), the calculated concentra-
tions of the species formed versus calcium concentra-
tion indicate that, while (3)2M and (3)3M are the main spe-
cies at low concentration, 3M2 and 3M3 are the major
compounds at high calcium concentration. The (3)2M and
(3)3M forms present the highest association constants,
whereas 3M3 has the smallest one. In comparison to the
alkyl-monosubstituted homologues under the same condi-
tions, the association constants of the (3)nM species are
higher by 1 or 2 orders of magnitude, which could be
explained by the involvement of the two arms of the
molecule in the association processes.
Figure 2. (top) 1H NMR chemical shift variations of the mentioned protons
of 3 (2.93 × 10-3 M) vs the Ca(CF3SO3)2 concentration in CD3CN at 293
K. Experimental values (2, (, b, O, 0, 4) and calculated (s) curves
obtained by fitting the data. See atom labeling in Scheme 2. (bottom)
Concentration of the species formed versus calcium concentration.
indicate that the conformational changes of the molecule
were less drastic than for the protonation reaction, the most
important variations being obtained for the CHb and CHh
groups. The CO groups of 3 were more perturbed upon Ca2+
addition than upon protonation. Thus, compound 3 behaved
differently upon protonation and calcium addition. In the
latter case, the shifts observed suggested that the CO groups
were the key points of the interaction. These observations
are reminiscent of the NMR behavior of compound 1 and
of that of the N-alkyl-monosubstituted homologues, upon
calcium addition.11,15
B. Analysis of the 1H NMR Data and Proposition of a
Model for Ca2+ Interaction. To get a better insight into
the 3-Ca2+ interaction process, the number and stoichiom-
etry of the calcium adducts involved were determined by
The formation of substoichiometric compounds of the LnM
type where n ) 2 or 3 is very likely, as the Ca2+ may interact
with six to nine donor atoms in the case of organic ligands
exhibiting the -CdC-CdO linkage, according to the
Cambridge Crystallographic Data Centre database. We have
previously developed this point in refs 15 and 19. The
formation of the 3M3 species may be surprising. However,
the triflate ion may also interact with the Ca2+ ion, for
example through its O and F atoms, thereby increasing the
number and nature of possible donor atoms and existing
1
processing the H NMR data acquired at 293 K. Actually,
the continuous shifts of the sharp peaks observed during the
calcium titration experiments were indicative of the presence
of fast equilibriums on the NMR time-scale,16 and the
chemical shift variations of six protons of 3 were successfully
plotted versus calcium concentration (Figure 2, dots). Figure
2 clearly illustrates that all these protons are involved in the
CO-centered interaction: the whole unsaturated core of the
(14) Silverstein, R. M.; Bassler, G. C.; Morril, T. C. In Spectrometric
Identification of Organic Compounds, 4th ed.; J. Wiley and Sons: New
York, 1981; Chapter 3, p 198.
(15) Maynadie´, J.; Delavaux-Nicot, B.; Lavabre, D.; Donnadieu, B.; Daran,
J. C.; Sournia-Saquet, A. Inorg. Chem. 2004, 43, 2064-2077.
(16) (a) Hynes, M. J. J. Chem. Soc., Dalton Trans. 1993, 311-312. (b)
Fielding, L. Tetrahedron 2000, 56, 6151-6170.
(17) Marcotte, N.; Fery-Forgues, S.; Lavabre, D.; Marguet, S.; Pivovarenko,
V. G. J. Phys. Chem. 1999, A103, 3163-3170.
(18) Fery-Forgues, S.; Lavabre, D.; Rochal, D. New J. Chem. 1998, 22,
1531-1538.
(19) Delavaux-Nicot, B.; Maynadie´, J.; Lavabre, D.; Lepetit, C.; Donnadieu,
B. Eur. J. Inorg. Chem. 2005, 2493-2505.
5694 Inorganic Chemistry, Vol. 45, No. 14, 2006

Chemosensor with Remarkable Fluorescence
Table 3. Assigment of the Main Peaks Obtained by Mass Spectrometry Using the Positive FAB and the ESI-MS* Techniques for Compound 3 (5 ×
10-3 M) in the Presence of Calciuma
[3MX]+
[(3)2M]2+
*
[(3)2MX]+
[(3)3M]2+
*
[(3)3MX]+*
[3M2X3]+
[3M3X5]+
Ca(CF3SO3)2
Ca(ClO4)2
773
723
604.5
604.5
1358
1307
896.5
896.5
1111.0
1449.0
1893
-
-
a M ) Ca2+; X ) CF3SO3 or ClO4
.
teristics of 3 are very similar to those of compound 1.
However, insertion of a supplementary -CdC- unit after
the CO group of 1 induces an increase of the electron density
in the organic moiety and, consequently, a decrease of its
oxidation potential of 90 mV. The same kind of observation
was made when comparing the monosubstituted derivative
[(C5H5)Fe(C5H4CO(CHdCH)2C6H4NMe2)] with compound
2.15 In reduction, the single wave observed at Ep ) -1.80
a
V for compound 3 was attributed to a reduction process
mainly located on the CO function.
Figure 3. Cyclic voltammograms of compounds 1 (- - -) and 3 (s).
Experimental conditions: Pt electrode (1 mm diameter) in 0.1 M solution
of nBu4NBF4 in CH3CN, scan rate 100 mV s-1, ligand concentration 10-3
M, reference electrode SCE.
Upon addition of 2 equiv of HBF4‚Et2O, the electrochemi-
cal solution of compound 3 in CH3CN turned from red to
red-pink. A cathodic shift of - 60 mV was induced for the
FeII/FeIII couple. This oxidation process became irreversible,
and a passivation phenomenon caused by a nonconducting
deposit on the electrode prevented determination of its E1/2
value by linear voltammetry. It is noteworthy that this
cathodic variation is weaker than that observed for compound
1 (-120 mV) under the same conditions. In parallel, the
complex wave attributed to the oxidation of the organic
moiety disappeared, and the wave corresponding to the NH+
reduction process appeared at -0.35 V. The final intensity
of these different processes is in agreement with the
formation of the protonated species, 4. Direct measurements
of the electrochemical characteristics of compound 4 were
not easy: under these conditions, a partial deprotonation
rapidly occurred and prevented accurate data from being
obtained.
interactions.20 This hypothesis could help in the understand-
ing of the formation of the LMm species.
Thus, for the concentration used, several species compete
in solution. It can be noted that, in the reported equilibriums,
weak intermolecular interactions21 (for example with adjacent
phenyl groups) or electrostatic interactions, rather than
classical complexation reactions, may also be considered.
In addition, samples of 3 (5 × 10-3 M) containing 0.5, 1,
and 2 equiv of calcium salt were prepared as for the NMR
measurements, and their mass spectra were recorded. The
positive FAB technique using an MNBA matrix and the ESI-
MS revealed all the peaks expected for the five compounds
(Table 3), providing thus strong support for their formation
in solution.
B. From Calcium-Ligand Interaction to Electrochemi-
cal Detection. To investigate the capacity of compound 3
as a calcium electrochemical sensor, 2 equiv of Ca(CF3SO3)2
was added to a solution of 3 under the same electrochemical
conditions. A broad ill-defined process indicated that the CO
reduction process was strongly perturbed, and a new ir-
reversible reduction process appeared around -0.35 V. In
oxidation, electrochemical characteristics close to those
induced by the in situ protonation of compound 3 were
obtained (see Figure 4). Further addition of salt (up to 10
equiv) had no significant effect.
Thus it seems that the interaction with calcium leads to
the same electrochemical response as protonation. Taking
into account the outcome of the electrochemical detection
of calcium by the monosubstituted homologues, we had
suspected a peculiar role of the electrolyte. Therefore, the
calcium-ligand interaction process was re-examined in the
presence of the Et4NBF4 supporting electrolyte, following
Electrochemical Studies. A. Characterization of Com-
pound 3 and In Situ Formation of 4. The electrochemical
properties of compounds 3 and 4 have been investigated in
CH3CN. A typical voltammogram of compound 3 is shown
in Figure 3. In oxidation, the wave observed in cyclic
voltammograms (CV) at Ep = 1.12 V is from the oxidation
a
of the ferrocene moiety and corresponds to a quasi-reversible
process. The first complex irreversible oxidation process of
compound 3, whose E1/2 value is 0.56 V, may be attributed
to the oxidation of the organic amine moiety.15,22 This
potential lies in the range of values reported for the oxidation
of some aza ferrocenyl compounds.23 This oxidation process
is reminiscent of the electrochemical behavior of the organic
compound CO(CHdCHC6H4NEt2)2 under the same condi-
tions.15 As shown in Figure 3, the electrochemical charac-
(20) (a) Lawrance, G. A. Chem. ReV. 1986, 86, 17-33. (b) Frankland, A.
D.; Hitchcok, P. B.; Lappert, M. F.; Lawless, G. A. J. Chem. Soc.,
Chem. Commun. 1994, 2435-2436. (c) Onoda, A.; Yamada, Y.; Doi,
M.; Okamura, T.; Ueyama, N. Inorg. Chem. 2001, 40, 516-521.
(21) (a) Gokel, G. W.; De Wall, S. L.; Meadous, E. S. Eur. J. Org. Chem.
2000, 2967-2978. (b) Meadous, E. S.; De Wall, S. L.; Barbour, L.
J.; Gokel, G. W. J. Am. Chem. Soc. 2001, 123, 3092-3107. (c) Ma,
J. C.; Dougherty, D. A. Chem. ReV. 1997, 13033-1324. (d) Futterer,
T.; Merz, A.; Lex, J. Angew. Chem., Int. Ed. Engl. 1997, 36, 611-
613.
(22) (a) Spescha, M.; Duffy, N. W.; Robinson, B. H.; Simpson, J.
Organometallics 1994, 13, 4895-4904. (b) Duffy, N. W.; Harper, J.;
Ramani, P.; Ranatunge-Bandarage, R.; Robinson, B. H.; Simpson, J.
J. Organomet. Chem. 1998, 564, 125-131.
(23) Andrews, M. P.; Blackburn, C.; McAleer, J. F.; Patel, V. D. J. Chem.
Soc., Chem. Commun. 1987, 1122-1124.
Inorganic Chemistry, Vol. 45, No. 14, 2006 5695

Delavaux-Nicot et al.
that the FeII/FeIII oxidation process becomes clearly irrevers-
ible for ligand 3.
The complete electrochemical study on compounds 3 and
4 allows us to strengthen our explanation for the intriguing
and uncommon cathodic potential shift observed upon cation
addition. It also suggests us that the singular behavior of
compound 3 upon cation addition is of the same nature as
that previously encountered with the first disubstituted
compound, 1, which still needed to be elucidated.11
It should be added that the electrochemical behavior of
compound 3 was also evaluated toward other cations. With
Li+, Na+, and K+, electrochemical detection was inefficient,
probably because weak associations were involved. The
copper and magnesium salts resulted in unclear electro-
chemical detection and were not studied further. Considering
these results and the fact that compound 3 presents an optimal
response with calcium, we have chosen to use this salt
throughout this work.
Figure 4. Cyclic voltammograms of compound 3 in oxidation before
(- - -) and after (s) addition of 2 equiv of calcium triflate. Experimental
n
conditions: Pt electrode (1 mm diameter) in 0.1 M solution of Bu4NBF4
in CH3CN, scan rate 100 mV s-1, ligand concentration 10-3 M, reference
electrode SCE.
the procedure detailed in ref 15. It was verified that the
calcium-ligand 3 interaction process was stable over 48 h
in the absence of the supporting electrolyte and that ligand
3 alone was insensitive to the presence of this ammonium
salt. In contrast, the preparation of a 1:2:2 mixture of 3/Ca2+/
Et4NBF4 in CD3CN clearly afforded complex 4. Therefore,
it appears that the Ca2+/BF4- couple was responsible for the
formation of the protonated species, 4, in the presence of a
small amount of water provided by the electrochemical or
NMR medium. The protonation that follows the Ca2+
interaction process with 3 allows indirect calcium detection.
Such a mechanism has already been encountered for the
monosubstituted compound 2 and its derivatives.15,19
Optical Detection of Calcium by Ligand 3. A. Study
by UV-vis Absorption Spectroscopy. As mentioned in the
Introduction, few ferrocenyl compounds have been investi-
gated for their ability to detect ions by both electrochemical
and optical methods.
The optical detection capabilities of ligand 3 toward
calcium were evaluated to give further insight into this
interaction process under new experimental conditions. UV-
vis absorption spectroscopy experiments were performed in
CH3CN. The absorption properties of compound 3 were first
determined. At a 2.7 × 10-5 M concentration in acetonitrile,
this compound gave an orange solution. The UV-vis
absorption spectrum exhibited an intense long-wavelength
band at λmax ) 420 nm with a molecular absorption
coefficient, ꢀ420, of 35 100 L mol-1 cm-1. This band was
attributed to a charge-transfer (CT) transition resulting from
the displacement of the electron density from the donor
amino group toward the acceptor carbonyl group.5b It was
only slightly red-shifted with respect to the CT band of
compound 1 (416 nm, ꢀ416 ) 44300 L mol-1 cm-1), but it
was significantly wider (full width at half-maximum ) 5580
and 4650 cm-1 for 3 and 1, respectively). When compared
to 1, the absorption spectrum also displayed a new weak
band peaking around 296 nm. These features can be
explained by the extension of the conjugated system resulting
from the presence of an additional ethylenic unit.25
Actually, the apparent cathodic shift of the iron potential
(∆Ep ) - 60 mV) observed for ligand 3 upon Ca2+ addition
a
corresponds to the shift induced by ligand protonation. Thus,
the observed ∆Ep of the FeII/FeIII couple represents the
a
difference of the oxidation potential between the protonated
form, [FeIIN2H22+] 4, and its neutral form, [FeIIN2] 3. In
compound 3, the first oxidation potential of the molecule,
attributed to the organic amino moiety, precedes the iron
oxidation potential. Consequently, the oxidation of 3 first
leads to the formation of a cationic radical species probably
[FeIIN22+•], whose Ep (Fe) value (1.11 V) is higher than that
a
observed for its corresponding [FeIIN2H22+] protonated form
(1.05 V). The calculated ∆Ep is negative, and the shift
a
observed is cathodic.
So the “unexpected cathodic shift” noticed for the iron
couple upon cation addition is connected with the uncommon
nature of molecule 3. In contrast, for classical electrochemical
ferrocenyl sensors whose first oxidation potential is from
the iron couple in the protonated and neutral forms, a positive
shift is observed upon cation addition.1b
In our comparison of the electrochemical behaviors of
compounds 3 and 1 upon calcium addition, two main
differences appear: first, increasing the distance between the
Our aim, in this case, was to explore a wide range of
concentration in calcium, which means that a high metal-
to-ligand ratio should be attained. In these conditions, the
use of calcium triflate did not allow accurate determination
of the binding constants, probably because of a side reaction
that occurs. Actually, working with the monosubstituted
compounds, we have recently shown that the use of the Ca-
(CF3SO3)2 salt at high concentration (>5 × 10-3 M) may
lead to total protonation of the chemosensor (10-5 M),
redox center and the amino site decreases the ∆Ep value of
a
the Fe oxidation process from -120 to -60 mV. For ligand
3, the electronic communication through the conjugated link
is lower than in 1, which is in agreement with the literature
data.24 Another main difference between ligands 1 and 3 is
(24) Plenio, H.; Yang, J.; Diodone, R.; Heinze, J. Inorg. Chem. 1994, 33,
4098-4104.
(25) Kawamata, J.; Akiba, M.; Inagaki, Y. Jpn. J. Appl. Phys. 2003, 42,
L17-L19.
5696 Inorganic Chemistry, Vol. 45, No. 14, 2006

Chemosensor with Remarkable Fluorescence
Figure 5. Absorption spectra of compound 3 (1.66 × 10-5 M) in
acetonitrile before (-) and after the addition of calcium perchlorate. Ca-
(ClO4)2‚4H2O from top to bottom at 418 nm: 0, 7.5 × 10-6, 5 × 10-5, 2.5
× 10-3, 1 × 10-2, 2.5 × 10-2, 5 × 10-2, 7.5 × 10-2, and 1.0 × 10-1 M.
whereas this phenomenon was not detected using the
perchlorate salt.26 This is the reason why we cannot propose
a model with the triflate salt under UV-vis conditions with
compound 3.
Consequently, the use of calcium perchlorate was pre-
ferred, this salt being classically used for this type of
spectroscopic study. The addition of calcium perchlorate to
a solution of 3 led to a reddening of the solution. The
intensity of the CT band regularly decreased, whereas the
λmax moved to 456 nm (Figure 5), with the formation of a
quasi-isosbestic point. The appearance of the band at long
wavelengths can be assigned to the interaction of the carbonyl
groups with the cation.11,27
Figure 6. Processing of the UV-vis absorption data for compound 3 (1.66
×10-5 M) in the presence of calcium perchlorate in CH3CN. (top)
Absorbance vs calcium concentration at different wavelengths in nanometers.
The points are experimental and the curves (lines) were calculated by fitting
the data. (bottom) Corresponding calculated concentrations of the different
species.
As for the NMR experiments, one of the aim of the UV-
vis absorption study was to determine the number of species
involved into the Ca2+-ligand interaction process in these
new conditions. To do so, the absorbance variation was
analyzed versus cation concentration. Absorbance was
recorded at six different wavelengths chosen to obtain
maximum information, and the absorption spectroscopic data
were processed according to the global curve-fitting method
described elsewhere (Figure 6 top).17,18 Among the different
models investigated, good fits were obtained by taking into
account the existence of only three species of different
stoichiometries: 3M, (3)2M, and (3)3M. The corresponding
association constants were 1.07 ×103, 6.40 ×106, and 3.82
×106 M-1, respectively (values given with a (15% error),
pointing out the stability of the substoichiometric compounds.
Calculated concentrations of the species versus cation
concentration are presented in Figure 6 (bottom).
However, according to the model proposed, it can be
calculated that for a ligand concentration of 1 × 10-3
M
and a calcium concentration of around 3.4 ×10-4 M, the
formation of the (3)3M species is favored with respect to
that of the (3)2M species, while the 3M species should be in
minor proportion. Such a mixture was prepared and directly
analyzed by electrospray mass technique. Actually, the
spectra clearly displayed peaks which can be attributed to
the (3)3M and (3)2M species as their (3)2Ca2+ and (3)3Ca-
(ClO4)+ forms. The calculated and experimental spectra of
the (3)3Ca2+ cation were in perfect agreement (see Supporting
Information S1). Moreover, although a small peak situated
at 312 could be attributed to the 3M2+ species, no 3Mn
species were identified. One possible formula for the (3)3M
compound is the coordination of the same calcium atom to
six carbonyl groups belonging to three different ligands.
Theoretical investigations performed with compound 1 show
that this hypothesis is conceivable.28
In addition to these titration experiments, mass spectrom-
etry measurements were performed with calcium perchlorate
and ligand 3 under the same experimental conditions as in a
previous section, and they revealed all the peaks expected
for the three species (see Table 3).Unfortunately, it was not
possible to obtain any X-ray structures for one of these
adducts.
When comparing the results obtained by UV-vis absorp-
tion spectroscopy using calcium perchlorate with those of
the NMR analysis carried out with calcium triflate, it appears
that the number of species involved in the ligand-calcium
interaction and the value of their association constants are
markedly different. This is probably caused by the nature of
the symmetrical ClO4- anion which allows fewer associations
(26) Maynadie´, J.; Delavaux-Nicot, B.; Lavabre, D.; Fery-Forgues, S. J.
Organomet. Chem. 2006, 1101-1109.
(27) Ukai, T.; Kawazura, H.; Ishii, Y.; Bonnet, J. J.; Ibers, J. A. J.
Organomet. Chem. 1974, 65, 253-266.
-
with calcium than the large CF3SO3 anion, bearing more
(28) Delavaux-Nicot, B.; Lepetit, C. Unpublished results.
Inorganic Chemistry, Vol. 45, No. 14, 2006 5697

Delavaux-Nicot et al.
Figure 7. Fluorescence spectra of ligand 3 (1.68 × 10-6 M) in acetonitrile before and after addition of calcium perchlorate (λex ) 452 nm). Ca(ClO4)2‚
4H2O: (a) from bottom to top 0, 1.0 × 10-6, 2.5 × 10-6, and 7.5 × 10-6 M; (b) from top to bottom 7.5 × 10-6, 1.0 × 10-5, 5.0 × 10-5, 1.0 × 10-4, 2.9
× 10-4, and 5.0 × 10-4 M; (c) from bottom to top 5 × 10-4, 7.5 × 10-4, 1.0 × 10-3, and 2.5 × 10-3 M; (d) from top to bottom 2.5 × 10-3, 5.0 × 10-3
,
7.5 × 10-3, 1.0 × 10-2, 2.5 × 10-2, 5.0 × 10-2, 7.5 × 10-2, and 10-1 M.
donating heteroatoms, as previously evoked. As we will show
it in a future paper, this seems to be a general rule when
using other disubstituted ligands of this family. However,
an NMR study repeated with calcium perchlorate showed
that the same interacting sites of ligand 3 were involved,
regardless of the anion used.
emission is not systematic for other symmetrical disubstituted
compounds of this family, as will be described in a future
study.
Compound 3 was studied in different solvents. Both
absorption and emission spectra were shifted to the red when
the solvent polarity increased. This behavior is reminiscent
of that of compound 1 and indicates that the dipole moment
of 3 also increases in the excited state. However, the
magnitude of the solvatochromic effect observed here must
be pointed out. For instance, the emission spectrum of 3 was
shifted from 536 to 630 nm when passing from dioxane to
acetonitrile, while the absorption spectrum was only shifted
by 4 nm. This leads to exceptionally high Stokes shifts,30
reaching more than 8000 cm-1 in acetonitrile. This effect is
generally correlated to a large change in geometry or to a
strong variation of the charges within the molecule upon
excitation, leading to the subsequent modification of the
solvent shell around the excited chromophore.30 In addition,
in a protic solvent, for example, ethanol, the absorption and
emission spectra were situated at longer wavelengths (λabs
) 434 nm, λem ) 662 nm) than in aprotic solvents of similar
polarity. This reveals the formation of a hydrogen bond
between the solvent molecules and the ligand carbonyl
group.31
B. An Intriguing Fluorescence Behavior upon Calcium
Addition. As already mentioned, very few ferrocenyl deriva-
tives have a satisfying fluorescence efficiency. Compound
3 was designed to possess this original property, and possibly,
to act as a fluorescent probe for cation detection.
In the absence of salt, the emission spectrum of 3 in
acetonitrile displayed an intense unresolved emission band
peaking at 628 nm, independent of the excitation wavelength.
It can be noted that the emission spectrum of 3 was red-
shifted by 68 nm when compared to that of compound 1.
This red shift can be assigned to the lengthening of the
conjugated π-electron system and may be of interest for
biological purposes.29 In CH3CN, the quantum yield was up
to 1.1 × 10-1 (i.e., slightly higher than that of compound 1,
7.7 × 10-2). This result was rather surprising. In fact, a
decrease of the quantum yield could have been expected
because the insertion of a CdC unit into the basic structure
of compound 1 may generate new rotation and vibration
modes, thus increasing the possible pathways for nonradiative
deactivation. It must also be underlined that the fluorescence
In the presence of calcium perchlorate, compound 3 (1.68
× 10-6 M) was excited at 452 nm to minimize the effect of
absorption variations. Figure 7 shows that an intriguing
(30) Suppan, P.; Ghoneim, N. In SolVatochromism; The Royal Society of
Chemistry: Cambridge, U.K., 1997; pp 17-18.
(31) Marcotte, N.; Fery-Forgues, S. Perkin Trans. 2 2000, No. 8, 1711-
1716.
(29) Burdette, S. C.; Walkup, G. K.; Spingler, B.; Tsien, R. Y.; Lippard,
S. J. J. Am. Chem. Soc. 2001, 123, 7831-7841. (b) Czarnik, A. W.
Chem. Biol. 1995, 2, 423-428.
5698 Inorganic Chemistry, Vol. 45, No. 14, 2006

Chemosensor with Remarkable Fluorescence
The emitting species is now different from the free ligand.
With the complexity of the system, it is very difficult to
identify this emitting species, although it is tempting to think
that the fluorescence comes from the 3M complex, which
forms in a large proportion at high calcium concentrations.
An interesting hypothesis would be that the fluorescence
revival observed is the result of the rigidification of the ligand
in the complex formed. A possible explanation for the final
collapse of fluorescence intensity (calcium concentration >
2.3 × 10-3 M) is that it comes from the presence of water
in solution, each calcium perchlorate molecule being associ-
ated with four molecules of water. The emissive complex
could coexist in solution with its hydrated form, the latter
being poorly fluorescent. For the sake of comparison, the
emission spectrum of 3 (1.44 × 10-6 M) in the absence of
calcium was recorded upon the addition of water. When the
water concentration was higher than 5.0 × 10-3 M, a regular
decrease of intensity, together with a red-shift, were observed.
For instance, with 5 × 10-1 M water, the fluorescence
intensity was reduced by 45%, and the spectrum was shifted
by 18 nm. This indicates that the excited state of 3 is affected
by the presence of water. However it can be noted that quite
high concentrations of water are necessary for a quenching
effect to be noticeable. It must be outlined that the sensitivity
to the presence of water only concerns the excited state of
3. It was not detected with the ground state of 3. For this
reason, a good fit of the UV-vis absorption data could be
attained without taking into account the involvement of water
in the complexation process.
To have a better insight into this strange phenomenon,
complementary experiments were performed. The concentra-
tion of each species present in solution was calculated, taking
into account the ligand and calcium concentrations used in
fluorescence experiments. To do so, the curve-fitting model
obtained when processing the UV-vis data was used, so
the calculated concentrations refer to the species in their
ground states. This approach confirms that the different
species appear in the same order as for UV-vis absorption
experiments. It was also noted that, at the point of maximum
quenching effect, the concentrations of the three species 3,
33M, and 32M are almost identical (S4). Let us assume that
the adducts 33M and 32M are not fluorescent and that the
only fluorescent species in solution is the free ligand, since
the concentration of the latter is divided by about three when
compared to its initial concentration, the fluorescence
intensity at this point should be one-third of the initial
fluorescence intensity. However, it is far lower, which
suggests that a more complex phenomenon, for example an
intermolecular quenching reaction, occurs in the excited state.
It is noteworthy that, according to this model, the subsequent
enhancement of fluorescence intensity corresponds to the
formation of the LM species.
Figure 8. Ligand 3 (1.68 × 10-6 M), variation of the fluorescence intensity
ratio at 630 nm versus calcium concentration in CH3CN (λex ) 452 nm).
“multistep behavior” was observed. First, the emission
intensity was very slightly increased until the Ca2+ concen-
tration reached 7.5 × 10-6 M, then it decreased steeply
without any shift in wavelength (630 nm) until fluorescence
was totally quenched at a 5 × 10-4 M salt concentration.
The addition of salt until a concentration of 2.5 × 10-3
M
led to a revival of fluorescence, the band now being centered
at 636 nm. Further stepwise addition of salt induced a new
decrease of the emission intensity, accompanied by a
subsequent red shift (12 nm) of the emission maximum.
Figure 8 illustrates this intriguing behavior at λem ) 630 nm.
The excitation spectrum was recorded for the free ligand
and in the presence of six different salt concentrations. For
the free ligand, the excitation maximum was situated at 424
nm, independent of the emission wavelength at which the
spectrum was recorded. This indicates that fluorescence
comes from only one species, as could be expected. It can
be noted that the excitation maximum was slightly red-shifted
by 4 nm with respect to the absorption maximum. This could
be a result of the free dye taking different conformations in
solution, some of them being nonemissive.30 For calcium
concentrations of less than 2.5 × 10-3 M, the excitation
spectrum was unchanged. For very high concentrations (1
× 10-1 M), the excitation spectrum moved to long wave-
lengths. The excitation spectrum peaked at 430 nm when
emission was recorded at 648 nm and at 438 nm when
emission was set at 700 nm. Conversely, the emission
spectrum passed from 648 to 694 nm when excitation was
performed at 452 and 500 nm, respectively. These observa-
tions indicate that at least two species responsible for
fluorescence coexist in solution, at high salt concentrations.
These results give a comprehensive view of the fluores-
cence behavior of compound 3 in the presence of calcium.
It seems clear that, in low salt concentrations, the emissive
species is the free ligand. The first collapse of the fluores-
cence intensity can be attributed to the fact that the free
species becomes scarce because of the formation of the
complexes. This implies that the first complexes formed
(most likely, the (3)3M and, maybe, the (3)2M species) are
not emissive. Then, when the fluorescence revival is observed
for calcium concentrations between 5 × 10-4 and 2.3 × 10-3
M, it is accompanied by a shift in the position of the emission
spectrum, even if the excitation spectrum was unchanged.
The titration experiment was repeated with two different
concentrations of the ligand (i.e., 7.4 × 10-7 and 3.9 × 10-7
M). The minima of the fluorescence titration curves were
then shifted, and their shapes were different from that of
the curve previously obtained with a ligand concentration
of 1.68 × 10-6 M. In particular, no total fluorescence
Inorganic Chemistry, Vol. 45, No. 14, 2006 5699

Delavaux-Nicot et al.
Scheme 3. Species Involved in the Equilibria Observeda
electrochemical properties, a noticeable E1/2 shift of the FeII/
FeIII couple was observed upon calcium triflate addition, but
this couple became irreversible. Indeed, we have rationalized
this efficient calcium sensing in terms of successive calcium
interaction and protonation reactions. Finally, in contrast to
similar compounds reported in the literature, our new
disubstituted ligand 3 presents good fluorescence properties
with an interesting quantum yield and a spectacular solva-
tochromic effect. To our knowledge, this is the first example
of a ferrocene-based sensor whose emission response shows
a “multistep” behavior in the presence of a cation. These
peculiar properties are directly linked to the insertion of an
additional -CHdCH- unit into the molecular framework
of 1. Surprisingly, under our experimental conditions, while
this insertion only induces modest changes in the electro-
chemical calcium detection, the fluorescence calcium sensing
is considerably modified and of an unprecedented and
intriguing nature. To our knowledge, this optical behavior
has never been observed with commercially available calcium
fluorescent sensors, which are, in fact, based on organic
derivatives generally used under physiological conditions.1i,34
In conclusion, we have demonstrated that the conjugated
purely ferrocenyl fluorescent compound 3 shows an original
behavior in both electrochemical and optical calcium sensing.
Our results support clearly the premise that other interesting
properties will also be obtained with simple but efficient
modifications of the basic framework of 1. In this respect,
we are studying new structures, bearing for example crown-
ether groups instead of terminating N-alkyl groups, to
improve sensitivity of the system. The possibility of develop-
ing molecules that can be used in aqueous media is also under
consideration.
a All the species are detected under NMR conditions with the Ca(CF3SO3)2
salt. Only the species written horizontally are detected under UV-vis
conditions with the Ca(ClO4)2 salt. As mentioned in the text and developed
in refs 15 and 19, one stoichiometry may correspond to several formulas.
quenching was observed. This can be explained by the fact
that the respective concentrations of all the species present
in solution varies depending on the initial concentration of
free ligand. Consequently, for a given calcium concentration,
the signal intensity can be somewhat different, according to
the initial ligand concentration.
Because compound 3 displays the same fluorescence
intensity in the presence of two different salt concentrations,
this compound does not appear as a convenient fluorescence
calcium sensor. However, the originality of its fluorescence
behavior in the presence of calcium perchlorate must be
underlined. In this respect, compound 3 totally differs from
its disubstituted homologue 1, for which a more classical
behavior will soon be presented.32 To our knowledge, such
an emission behavior has never been reported previously for
organic chalcone derivatives in the presence of salts and is
also totally unprecedented for ferrocene compounds.
Concluding Remarks
We have designed compound 3 according to a “three-
component conjugated concept” and have shown that this
new ligand is easy to prepare and allows calcium detection
by both electrochemistry and UV-vis absorption techniques.
This multiresponsive ferrocene receptor binds a calcium guest
via a complex process involving its CO functions and its
whole unsaturated core. In addition, we have proposed the
quantification of the ligand-cation interaction process for
calcium triflate under NMR conditions and for calcium
perchlorate under UV-vis absorption spectroscopy condi-
tions. In each case, for the ligand-calcium equilibrium
considered, several species were involved as summarized in
Scheme 3. Furthermore, their existence was supported by
mass spectrometry, and we have successfully determined the
values of their association constants. The use of two different
calcium salts certainly emphasized the role of the anion,
which can determine the number and nature of the formed
adducts. However, this phenomenon may also result from
the different concentrations and time scales employed in the
different techniques.33 At this point, we would like to
underline the fact that the subject of ligand-ion interaction
has seldom been addressed in previous publications dealing
with electrochemical ferrocenyl ion sensors. In regard to the
From a general point of view, this study suggests that the
development of this kind of receptors opens wide perspec-
tives in the design of a novel generation of attractive
multiresponsive molecules with surprising properties.
Experimental Section
Materials. Toluene, THF, and ether were distilled over sodium/
benzophenone; pentane, dichloromethane, and CH3CN (pure SDS)
were distilled over CaH2 and stored under argon. Analytical grade
EtOH (purex SDS) was simply degassed. Fe(C5H4COMe)2 (95%),
CHOCHdCHC6H5NMe2 (98%), and HBF4 54% in Et2O were from
Aldrich. Calcium salts, Ca(CF3SO3)2 (96%) and Ca(ClO4)2‚4H2O
(99%), were from Strem and Aldrich, respectively. Caution:
Perchlorate salts are hazardous because of the possibility of
explosion!
General Instrumentation and Procedures. All syntheses were
performed under a nitrogen atmosphere using standard Schlenk tube
techniques. The solutions of compounds 3 and 4 were light-
protected before each measurement. IR spectra were recorded on a
Perkin-Elmer GX FT-IR spectrophotometer. Samples were run as
KBr pellets. Elemental analyses were carried out on a Perkin-Elmer
2400 B analyzer at the L.C.C. Microanalytical Laboratory in
Toulouse. Mass spectra were obtained at the Service Commun de
(32) Delavaux-Nicot, B.; Maynadie´, J. Unpublished results.
(34) For example, see: (a) Haugland, R. P. Handbook of Fluorescent Probes
and Research Products, 9th ed.; Molecular Probes, Invitrogen:
Carlsbad, CA, 2002; references therein. (b) Brownlee, C. Trends Cell
Biol. 2000, 10, 451-457.
(33) Beer, P. D.; Blackburn, H. S. C.; McAleer, J. F. J. Organomet. Chem.
1988, 350, C15-C19. (b) Beer, P. D.; Keefe, A. D.; Blackburn, H. S.
C.; McAleer, J. F. J. Chem. Soc., Dalton Trans. 1990, 3289-3294.
5700 Inorganic Chemistry, Vol. 45, No. 14, 2006

Chemosensor with Remarkable Fluorescence
Spectrometrie de Masse de l′Universite´ Paul Sabatier et du CNRS
de Toulouse. Fast-atom bombardment (FAB > 0) spectra were
performed on a Nermag R10-10H spectrometer. A 9 kV xenon atom
beam was used to desorb samples from the 3-nitrobenzyl alcohol
matrix. Other spectra were performed on a triple quadrupole mass
spectrometer (Perkin-Elmer Sciex API 365) using the electrospray
ionization mode. The infusion rate was 5 µL/min. 1H and 13C NMR
spectra were performed on Bruker AC 200, AM 250, DPX 300,
and AMX 400 spectrometers. 1H and 13C NMR spectra were
referenced to external tetramethylsilane. For 2D NMR experiments,
the observation frequencies were in the range of 400.13 MHz for
1H and 100.62 MHz for 13C.
Electrochemical Studies. Voltammetric measurements were
carried out with a homemade potentiostat35 using the interrupt
method to minimize the uncompensated resistance (iR drop).
Experiments were performed at room temperature in an airtight
three-electrode cell connected to a vacuum/argon line. The reference
electrode consisted of a saturated calomel electrode (SCE) separated
from the solution by a bridge compartment filled with the same
solvent and supporting electrolyte solution. The counter electrode
was a platinum wire of ca. 1 cm2 apparent surface. The working
electrode was a Pt electrode (1 mm diameter). The supporting
electrolyte nBu4NBF4 (99%) (Fluka electrochemical grade) was
melted and dried under vacuum for 1 h. All solutions measured
were 1.0 × 10-3 M in organometallic complex and 0.1 M in
supporting electrolyte. The solutions were degassed by bubbling
experimental values of the variation of the chemical shifts,
respectively. The number of unknown parameters includes the
equilibrium constant for each species and their chemical shifts. This
number may seem important, but it must be underlined that a
simultaneous fitting on several distinct signals, 6 in our case, is
much more difficult to process than a fit on a single signal.
Optical Measurements. Apparatus. UV-vis absorption spectra
were recorded on a Hewlett-Packard 8452A diode array spectro-
photometer. The absorption spectra did not vary over a period of
2 h. Cuvettes of 1 cm optical path length were used. Steady-state
fluorescence work was performed on a Photon Technology
International (PTI) Quanta Master 1 spectrofluorometer. All
fluorescence spectra were corrected. The fluorescence quantum yield
was determined relative to coumarin 6 in ethanol as the standard
(ΦF ) 0.78).36 The measurements were conducted at 20°C in a
thermostated cell.
Data Analysis. The absorption experimental data were processed
on a HP 9000 series 710 workstation. Absorbance, A, was related
to the concentration, Ci, using Beer-Lambert’s law, A ) l∑(ꢀiCi),
where ꢀi is the molar absorption coefficient of the species i and l
is the optical path length. The system of three equilibrium equations
with three independent variables was numerically solved by an
iterative method. The sum of the squares of the differences between
the experimental values and those of the numerical calculation was
minimized by a Powell nonlinear minimization algorithm. The
method has been extensively described in a previous paper.17
[Fe(C5H4CO(CHdCH)2C6H4NMe2)2] (3). A light-protected
mixture of Fe(C5H4COMe)2 (0.100 g, 0.37 × 10-3 mol), CHOCHd
CHC6H4NMe2 (0.130 g, 0.74×10-3mol), and 1 equiv of NaOH was
dissolved in ethanol (10 mL) and stirred for 4 h at room temperature.
The mixture was evaporated to dryness. The residue was purified
by column chromatography on dried silica (eluent ) 5:2 petroleum
ether/ethyl acetate), and the red phase was extracted with THF as
the eluent (3×). After evaporation of the solvent, the product was
washed with pentane (30 mL × 2) and dried to afford the desired
product as a deep orange powder in a 65% yield (m ) 0.140 g).
1H NMR (CD3CN, 293 K): δ 2.99 (s, 12H, Hp), 4.59 (br s, 4H,
3JHeHf ) 2.0 Hz, Hf), 4.84 (br s, 4H, 3JHeHf ) 2.0 Hz, He), 6.64 (d,
argon before the experiments. With the above reference, E1/2
)
0.45 V vs SCE was obtained for 1 mM ferrocene (estimated
experimental uncertainty of (10 mV). Cyclic voltammetry was
performed in the potential range of -2 to 2 V versus SCE scanning
from 0 toward 2 V/SCE for the oxidation studies (and from 0 toward
-2 V/SCE for the reduction studies) at 0.1 V s-1 at room
temperature. Before each measurement, the electrode was polished
with emery paper (Norton A621). To calculate the half wave
potential (E1/2), a quasi-steady-state behavior (at the Pt working
electrode, 1 mm in diameter) is obtained by the use of linear
voltammetry at 5 mV s-1. For cation detection experiments,
concentrated acetonitrile solutions of calcium triflate (0.3-10 equiv)
were syringed into the ferrocenyl solution under an argon atmo-
sphere, keeping the total volume of the electrochemical mixture
constant. The solution was immediately degassed and examined.
Proton NMR Titration Studies. Proton NMR titrations were
typically performed as follows. A solution (500 µL) of the receptor
3 in a deuterated solvent (10-2 M) was added (using a microsyringe)
into NMR tubes containing the appropriate quantities of solid Ca-
(CF3SO3)2 salt under an inert atmosphere, while the NMR spectrum
of the receptor was monitored. The samples of solid calcium were
prepared by evaporating the corresponding calculated volumes of
a calcium guest solution (10-2 M) in acetonitrile. Stability constants
were evaluated from titration data using the method indicated in
the main text.
3
3
4H, JHcHd ) 8.8 Hz, Hd), 6.67 (d, 2H, JHaHb ) 14.8 Hz, Ha),
6.88 (dd, 2H,3JHbHg ) 11.2 Hz, JHhHg ) 15.2 Hz, Hg), 7.02 (d,
3
2H, 3JHhHg ) 15.2 Hz, Hh), 7.36 (d, 4H,3JHcHd ) 8.8 Hz, Hc), 7.46
(dd, 2H, 3JHbHg ) 11.2 Hz, 3JHaHb ) 14.8 Hz, Hb). 13C {1H} NMR
(CD3CN, 293 K): δ 39.79 (CHp), 71.27 (CHe), 73.94 (CHf), 83.40
(Cipso-C5H4), 112.40 (CHd), 122.70 (CHg), 124.49 (Cipso-C),
124.49 (CHa), 129.06 (CHc), 142.61 (CHh), 142.65 (CHb), 151.57
(Cipso-N), 191.81 (CO). IR (CH3CN): 1522, 1559, 1570, 1604, 1645
(νCO), 2873-2998 (νCH). ESI-MS: 585.1 [M + H] +. Anal. Calcd
for 3, C36H36N2O2Fe: C, 73.97; H, 6.21; N, 4.79. Found: C, 73.72;
H, 6.19; N, 4.76.
Interaction of 3 with 2 equiv of Ca(CF3SO3)2. 1H NMR(CD3-
CN, 293 K): δ 2.99 (s, 12H, Hp), 4.80 (broad s, 4H, Hf), 5.00
Residual Error (Er) of the Curve-Fitting Model. The residual
error was calculated using the expression
(broad s, 4H, He), 6.55 (d, 4H, 3JHcHd ) 8.8 Hz, Hd), 6.55 (d, 2H,
3
3JHaHb ) 14.9 Hz, Ha), 6.78 (dd, 2H,3JHbHg ) 12.0 Hz, JHgHh
)
14.9 Hz, Hg), 7.08 (d, 2H, 3JHgHh ) 14.9 Hz, Hh), 7.22 (d, 4H,3JHcHd
) 8.8 Hz, Hc), 7.74 (dd, 2H, 3J HbHg ) 12.0 Hz, 3JHaHb ) 14.9 Hz,
Hb). 13C {1H} NMR (CD3CN, 293 K): δ 39.72 (CHp), 72.53
(CHe), 75.38 (CHf), 82.89 (Cipso-C5H4), 112.19 (CHd), 122.03
(CHg), 122.38 (CHa), 123.95 (Cipso-C), 129.87 (CHc), 147.10
(CHh), 148.76(CHb), 152.04 (Cipso-N), 194.05 (CO).
n
1
Er )
(ci - ei)2
∑
n - p - 1i)1
where n is the total number of points (6 × 10), p is the number of
unknown parameters (35 with five species, 21 when only three
species are taken into account), and ci and ei are the calculated and
[Fe(C5H4CO(CHdCH)2C6H4NHMe2)2] [BF4]2 (4). HBF4‚Et2O
(2 equiv) was slowly syringed into a stirred solution of 3 (0.08 ×
(35) Cassoux, P.; Dartiguepeyron, R.; de Montauzon, D.; Tommasino, J.
(36) Reynolds, G. A.; Drexhage, K. H. Opt. Commun. 1975, 13, 222-
B.; Fabre, P. L. Actual. Chim. 1994, 1, 49-55.
225.
Inorganic Chemistry, Vol. 45, No. 14, 2006 5701

Delavaux-Nicot et al.
10-3 M) in acetonitrile (8 mL). The light-protected mixture was
stirred for 1 h. After the solvent was evaporated, the product was
washed with ether (30 mL) and pentane (50 mL) and dried under
fellowship to J. M.). The reviewers are acknowledged for
constructive comments.
1
vacuum. A violet powder was obtained in a 68% yield. H NMR
Supporting Information Available: Figure S1 showing the
calculated (top) and experimental (bottom) ESI-MS spectra of the
(3)3Ca2+ cation in CH3CN, Figure S2 showing the 1H NMR
chemical shift variations of the mentioned protons of 3 versus the
Ca(CF3SO3)2 concentration in CD3CN at 293 K (the calculated
curves are obtained by fitting the data with the K1K2K4 model), S3
listing the calculated values of the ∆ (δ H, ppm) of 3 obtained
when using the K1-K5 model and the K1K2K4 model, and S4
showing the calculated concentration of each 3 adduct in the excited
state using the curve-fitting model obtained when processing the
UV-vis absorption data which refer to the ground state ([3] )
1.68 × 10-6 M in CH3CN). This material is available free of charge
via the Internet at http://pubs.acs.org.
(400 MHz, CD3CN, 293 K): δ 3.25 (s, 12H, Hp), 4.66 (t, 4H,
3
3JHeHf ) 2 Hz, Hf), 4.89 (t, 4H, JHeHf ) 2 Hz, He), 6.86 (d,
2H,3JHaHb) 14.8 Hz, Ha), 7.11 (d, 2H, 3JHgHh ) 15.6 Hz, Hh), 7.21
3
3
(dd, 2H, JHbHg ) 10.8 Hz, JHgHh ) 15.6 Hz, Hg), 7.46 (dd, 2H,
3JHbHg ) 10.8 Hz, 3JHaHb ) 14.8 Hz, Hb), 7.57 (d, 4H, 3JHcHd ) 8.8
Hz, Hd) 7.72 (d, 4H, JHcHd ) 8.8 Hz, Hc), 8.95 (sl, 2H, NH+).
3
13C{1H} NMR (100.6 MHz, CD3CN, 293 K): δ 47.41 (CHp), 71.53
(CHe), 74.75 (CHf), 82.57 (Cipso-C5H4), 121.60 (CHd), 128.19
(CHa), 129.15 (CHc), 130.36 (CHg), 138.33 (CHh), 139.16 (Cipso
-
C), 140.32 (CHb), 142.11 (Cipso-N), 191.85 (CO). MS (ESI, CH3-
CN): 673.5 [M - BF4-]. Anal. Calcd for 4, C36H38N2O2FeB2F8:
C, 56.88; H, 5.04; N, 3.68. Found: C, 56.52; H, 4.95; N, 3.75.
Acknowledgment. This work was supported by the
CNRS and the French Ministry of Research (doctoral
IC060535E
5702 Inorganic Chemistry, Vol. 45, No. 14, 2006