The synthesis and complexation of a cobaltocenium-based redox-active cryptand containing the phenanthroline unit

Article abstract of DOI:10.1016/S0022-328X(01)01415-2

The synthesis and characterization of a cobaltocenium-based redox-active cryptand 3 containing a phenanthroline unit is described. Complex formation (with Ca²⁺) was studied by ¹H/¹³C-NMR, and the effects of complexation on the redox properties of 3 were monitored by cyclic voltammetry. The fluorescence spectra of the complexes of 3 and its ferrocene analogue (4) with Eu³⁺ are also reported, and the low intensity of the fluorescence in each case attributed to carbonyl coordination of Eu³⁺ rather than metallocene quenching.

Full text of DOI:10.1016/S0022-328X(01)01415-2

Journal of Organometallic Chemistry 648 (2002) 8–13
www.elsevier.com/locate/jorganchem
The synthesis and complexation of a cobaltocenium-based
redox-active cryptand containing the phenanthroline unit
C. Dennis Hall *, Natasha Djedovic
Department of Chemistry, King’s College, Strand, London WC2R 2LS, UK
Received 23 January 2001; received in revised form 19 October 2001; accepted 22 October 2001
Abstract
The synthesis and characterization of a cobaltocenium-based redox-active cryptand 3 containing a phenanthroline unit is
1
described. Complex formation (with Ca2+) was studied by H/13C-NMR, and the effects of complexation on the redox properties
of 3 were monitored by cyclic voltammetry. The fluorescence spectra of the complexes of 3 and its ferrocene analogue (4) with
Eu3+ are also reported, and the low intensity of the fluorescence in each case attributed to carbonyl coordination of Eu3+ rather
than metallocene quenching. © 2002 Elsevier Science B.V. All rights reserved.
Keywords: Cobaltocenium; Redox-active cryptand; Phenanthroline unit
1. Introduction
2. Results and discussion
The synthesis, structure, electrochemistry and com-
plex formation of cryptands containing metallocene
units have attracted considerable attention in recent
years [1–3]. These redox-active compounds are poten-
tially useful as sensors for the detection and quantita-
tive estimation of metal cations by electrochemical
techniques [4] as catalysts in a variety of reactions [5]
and possibly as voltage-regulated optical switches [6,7].
With respect to the latter application, all the com-
pounds synthesized so far have contained ferrocene, a
potential fluorescence quencher, as the electroactive
unit. This paper describes the synthesis, complexation,
electrochemical and photochemical properties of an
analogous system containing the cobaltocenium unit.
Compound 3 was prepared in 26% isolated yield by the
condensation of the tetraazamacrocycle (1) with cobal-
tocenium diacid chloride (2); a route based on the
synthesis of the analogous ferrocene cryptand 4 [6].
2.1. NMR data
2.1.1. Cryptand 3
n.b. The numbering is not systematic but has been
chosen to coincide with that used in Ref. [6] in the
interpretation of NMR data.
It is clear from the complexity of the H- and 13C-
1
NMR spectra that the molecule is unsymmetrical which
means that in common with all compounds of this type
examined so far [6,8–11] the carbonyl groups are trans
to each other when observed on the NMR time scale.
In particular, the 13C spectrum consists of 32 separate
signals comprised of four NCH2 carbons, four OCH2
carbons, eight cobaltocenium CH carbons, two cobal-
tocenium (ipso) carbons, six phenanthroline CH car-
bons, six (ipso) phenanthroline carbons and two
1
carbonyl carbons. Likewise, the H-NMR reflects the
unsymmetrical structure by, for example, showing eight
cobaltocenium proton multiplets, two widely-spread
AB quartets for 7AB and 20AB, six phenanthroline pro-
tons (as two sets of double doublets and an AB quartet)
and complex patterns for the NCH2 and OCH2 protons
of the aliphatic chain. The application of COSY 45, the
* Corresponding author. Present address: Department of Chem-
istry, University of Florida, PO Box 117200, Gainesville, FL 32611,
USA. Fax: +1-352-392-8758.
1
E-mail address: cdennishall@aol.com (C.D. Hall).
DEPT technique and H/13C heteronuclear correlation
0022-328X/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved.
PII: S0022-328X(01)01415-2

C.D. Hall, N. Djedo6ic / Journal of Organometallic Chemistry 648 (2002) 8–13
9
spectroscopy, enabled an almost unambiguous assign-
ment of all the NMR data to be achieved (Table 1).
The keys to the assignment are as follows. In the
1H-NMR spectrum one cobaltocenium proton is at
unusually low field (7.85 ppm) and this is assumed to be
H-2 cis to the phenanthroline group and cis to C-6 thus
placing the proton in the anisotropic deshielding cone
of the carbonyl function. With this much assumed, the
COSY-45 and 1H/13C correlation spectra enable the
assignment of all the protons and carbons around the
cyclopentadiene rings. Secondly, if it is assumed, as
suggested by models, that H7B and H22B lie in the
deshielding zones of the C-6 carbonyl and a cyclopenta-
diene unit, respectively, then the two AB quartets can
be assigned completely. Finally, by assuming that the
carbon atoms trans to the carbonyl groups are at the
lowest fields in the 13C spectrum [12], C-7, C-20 and
C-22–C-27 can also be assigned. Thus the only slight
ambiguity lies in the assignment of protons 9–16 and
23–26 and in the case of the 13C data, the assignment
of 8–19. One is tempted to assume that carbon atoms,
8–12 and 19 are at lower field than their respective
counterparts (13–18) and that likewise protons 9, 10
and 12 are at a lower field than H-13, 15 and 16.
Although the assumption is a reasonable one, it should
be treated with caution since it may represent an over-
interpretation of the data.
2.1.2. Complex of 3 with Ca2+
1H and 13C solution NMR was used to study com-
plex formation of 3 with a number of cations. For
example, the addition of one equivalent of Ca2+ to a
l
solution of 3 caused a dramatic change in both the H
and 13C spectra of the host molecule. Thus the phenan-
throline protons were shifted slightly downfield and
were seen as two unequally populated sets of two
doublets and a singlet. The integrated ratio of each set
was ca. 3:2 with the higher field set predominating. This
is reminiscent of the ferrocene analogue [6] and is due
to the formation of a mixture of 2:1 and 1:1 host: Ca2+
complexes with the former predominating in a molar
ratio of 64:36. The low field cyclopentadiene proton of

10
C.D. Hall, N. Djedo6ic / Journal of Organometallic Chemistry 648 (2002) 8–13
Table 1
Each set consisted of 16 signals comprised of two
NCH2 carbons, two OCH2 carbons, four cobaltoce-
nium CH carbons, one ipso-cobaltocenium carbon,
three phenanthroline CH carbons, three-ipso-phenan-
throline carbons and one carbonyl carbon (Table 2b).
The data are consistent with the existence of a plane of
symmetry within both complexes which, in common
with previous studies, implies that the carbonyl groups
within the host molecules are now cis to each other and
in6ol6ed in the coordinati6e bonding with the Ca2+ ion.
This is emphasized by the observation that the carbonyl
carbon signals of the complexes are observed at 167.7
and 169 ppm, whereas the carbonyl groups of the host
appear at 164 and 165 ppm.
1H- and 13C-NMR data on cryptand 3
1H
13C
Assignment
l (ppm)
Assignment
l (ppm)
10/15
12/13
9/16
8.40 (d/d)
7.94 (ABq)
7.65 (d/d)
10/15
12/13
9/16
138.6; 137.9
128.1; 127.2
122.6, 124.1
158.5,157.7,147.0,
146.5
129.8, 129.4
165.1: 164.0
105.4; 97.4
88.9
8,17,18,19
11/14
6/21
1/1%
2
2
3
7.86
5.55
3
87.6
4
5.85
4
87.3
5
2%
6.00
6.45
5
2%
86.0
89.8
2.2. Cyclic 6oltammetry (CV)
3%
5.90
3%
88.9
4%
5%
5.90
6.12
4.44
5.96
4.43
5.13
2.89 (m)
4.77
3.40
4%
5%
7
87.7
88.1
55.0
Cyclic voltammetry of the ligand alone and in the
presence of Ca2+ was carried out in acetonitrile with
TBAP as the supporting electrolyte (Fig. 1). Clearly,
complexation with Ca2+ causes an anodic shift of 170
mV, but equally there was always a discernible amount
of free ligand available even at a four equivalent excess
of [Ca2+] over [ligand]. This is emphasized by the
deconvolution plots (dI1/dt and dI1/dE vs. E) (Fig. 2a
and b), which also show that the CV waves are only
quasi-reversible since the oxidation and reduction peaks
do not coincide but are dispersed approximately 10 mV
either side of the E1/2 value (at −690 mV). All three
CV traces show that in the case of the complexed
systems, the oxidation wave of the free ligand in the
mixture is at a much lower intensity than the reduction
wave which implies that in the reduced form of the
electroactive center (Co2+) the ligand binds Ca2+
strongly, but on oxidation the Ca2+ is, to some extent,
released prior to reduction. This conclusion is corrobo-
rated by the observation that the intensity of the reduc-
tion wave of the complex is less than that of the
oxidation wave. The CV as a whole is reminiscent of
that observed with the ferrocene analogue (4) and
Mg2+ [6], although the comparison is somewhat tenu-
ous, since the latter was necessarily acquired with pro-
pylene carbonate as solvent. With Ca2+ and 4, only an
‘‘averaged’’ wave was observed (with a maximum shift
of ca. 150 mV), which implies a weaker complex than
that observed with 3. Again the comparison has to be
viewed with caution in view of the different solvent
systems.
7A
7B
20A
20B
22A
22B
27A
27B
23AB
20
22
27
23
24
25
26
56.0
48.1
53.5
69.2
71.2
71.8
71.0
4.23
2.44 (A)
3.22 (B)
2.92 (A)
3.33 (B)
2.79 (A)
3.33 (B)
3.62 (A)
3.68 (B)
24AB
25AB
26AB
the host at 7.85 ppm was now found in the conven-
tional cyclopentadiene region (5.8–6.6 ppm) where two
sets of four cobaltocenium CH signals were observed
again in the integrated ratio of ca. 3:2. The two AB
quartets associated with the CH2 groups at C-7 and
C-20 were also observed in more conventional positions
(4.9/3.35 for the 2:1 complex and 4.5/3.55 for 1:1
stoichiometry) again in the same ratio of 3:2. The
remainder of the proton spectrum was inevitably rather
complex, but, in general, confirmed the analysis. Addi-
tion of a second equivalent of Ca2+ ion merely
changed the ratio of 2:1 vs. 1:1 complexes so that the
latter predominated in a ratio of 3:2 (i.e. 60% 1:1
complex) (Table 2a). The assignments are based on the
reasonable assumption that the lower field sets of pro-
tons are due to the 1:1 complex where the charge
density of the cation would be more pronounced. The
13C spectra of the Ca2+ complexes serve to confirm
these conclusions. At a molar ratio of 1:1, two un-
equally populated sets of 13C signals were observed.
2.3. Fluorescence spectra
Ultraviolet irradiation of complexes of 4 with Eu3+
(Fig. 3b), Tb3+or Yb3+ produced only weak fluores-
cence in the expected visible region of the spectrum.
The low level of fluorescence was ascribed either to

C.D. Hall, N. Djedo6ic / Journal of Organometallic Chemistry 648 (2002) 8–13
11
Table 2
Data on the 1:1 and 2:1 (host:guest) complexes formed by 3 and calcium triflate
2:1 (Host:Ca2+
)
1:1 (Host:Ca2+
)
l
Assignment
l
Assignment a
(a) 1H-NMR
8.40 (J=8.2)
7.86 (s)
7.62 (J=8.2)
6.45 (m)
H 10/15
H 12/13
H 9/16
cp
8.67 (J=8.2)
8.12 (s)
7.99 (J=8.2)
6.61 (m)
H 10/15
H 12/13
H 9/16
cp
6.30 (m)
cp
6.45 (m)
cp
5.93 (m)
cp
5.98 (m)
cp
5.87 (m)
cp
5.93 (m)
cp
4.91/3.99
4.00–1.35 (multiplets)
AB q (7/20)
OCH2 /NCH2 chain
5.87/4.503
AB q (7/20)
NCH2
OCH2 (23/26)
OCH2 (24/25)
3.99, 3.541 (m)
3.25, 2.454 (m)
2.801, 1.53 (m)
(b) 13C-NMR
52.7, 55.7
67.4, 69.7
86.3, 86.7, 87.7, 90.8,
99.4
2×NCH2
2×OCH2
4×CoCH
1/1%
55.7, 53.3
69.8, 67.3
91.1, 89.0, 86.8, 86.2
99.8
C7/20; C22,27
C24/25; C23,26
4×CoCH (5,2,4,3)
1/1%
125.6, 127.9, 140.9
129.1, 145.2, 160.8
167.7
phen CH
phen C
CꢀO
141.2, 128.1, 125.9
160.2, 144.9, 130.0
168.6
10/15; 12/13; 9/16
8,17,18,19; 11,14
C=O
a In the order shown against each chemical shift.
quenching by the ferrocene unit or to coordination of
the cations by the carbonyl oxygens, thus distancing the
fluorescent ion from the phenanthroline unit and hence
inhibiting energy transfer. A similar set of results was
obtained using the cobaltocenium cryptand 3 and hence
the reported spectra (Fig. 3a, b and Table 3) are
restricted to the results with Eu3+. The cobaltocenium
ligand (3) was designed to distinguish between quench-
ing by the metallocene unit and distancing by carbonyl
coordination and since the cobaltocenium unit, an 18-
electron system, would not be expected to quench
fluorescence, it is likely that the latter explanation is
Fig. 2. (a) Deconvolution data (dI1/dt) vs. E on the CV trace for the
complexation of 3 with four equivalents of Ca2+. (b) Deconvolution
data (dI1/dE) vs. E on the CV trace for the complexation of 3 with
Fig. 1. Cyclic voltammetry traces on 3 and its complexation with
four equivalents of Ca2+
.
Ca2+
.

12
C.D. Hall, N. Djedo6ic / Journal of Organometallic Chemistry 648 (2002) 8–13
(or 100.6) MHz for 13C-NMR. Chemical shifts (l) are
given in parts per million (ppm) downfield from Me4Si
as internal standard. 13C-DEPT (distortionless enhance-
ment of polarization transfer) spectra were recorded
using the 135° pulse sequence, giving positive signals
for CH3 and CH, negative signals for CH2 and no
signal for quaternary (ipso) carbon atoms. High resolu-
tion fast-atom-bombardment mass spectrometry
(HRMS-FAB) was performed at the ULIRS Mass
Spectrometry Facility, School of Pharmacy on a VG
ZAB-SE instrument using VG Opus software (resolu-
tion 8–10 K). Infrared spectra were recorded on a
Perkin–Elmer Paragon 1000 FT-IR spectrometer using
NaCl plates (neat) or KBr discs (KBr). Elemental
analyses were determined by the microanalytical service
at University College London. Thin layer chromato-
graphic (TLC) analyses were performed on aluminum
oxide 60 F254 neutral on aluminum sheets with a 0.2
mm layer thickness (Merck), using 5% MeOH–CH2Cl2
as eluent. Preparative chromatography columns were
packed with alumina [neutral, type 507C, Fluka, (0.05–
0.15 mm; pH 7.090.5)], or silica gel 60 (F254, 0.035–
0.070 mm, Merck). Solvents were dried and purified
,
prior to use and stored under nitrogen over class 4 A
molecular sieves (4–8 mesh). CH2Cl2, MeCN and Et3N
were all distilled under nitrogen from calcium hydride.
MeCN for electrochemistry was dried prior to use as
indicated above and tetrabutylammonium perchlorate
(TBAP) was purchased from Fluka and used without
further purification.
Cyclic voltammograms were recorded using an
EG&G Princeton Applied Research Versastat™ poten-
tiostat and the current-potential traces analyzed using
EG&G 270/250 Research Electrochemistry Software.
The data were recorded using either a platinum or a
glassy carbon (diameter, 0.3 cm) working electrode, a
platinum counter electrode and a Ag/AgCl reference
electrode (at 0.222 V). The working electrodes were
polished with alumina at frequent intervals. An MeCN
solution (containing 0.2 M TBAP as supporting elec-
trolyte) of the ligand (at ca. 2 mmol) was placed in the
unsealed one-compartment cell and degassed by bub-
bling with Ar prior to each recording. An Ar atmo-
sphere was maintained above the solution while the
experiments were in progress. Addition of calculated
quantities of calcium triflate solution was accompanied
by stirring, but recordings were made on the unstirred
solutions.
Fig. 3. (a) Luminescence spectra for the complexes of 3 with Eu3+
at 10 equivalents excess and 50 equivalents excess of Eu3+. (b)
Luminescence spectra for the complexes of 4 with Eu3+ at 50
equivalents excess of Eu3+
.
correct a conclusion which is reinforced by the
13C-NMR data on the complexes.
3. Experimental
1H-NMR spectra were recorded on a Bruker AM-360
spectrometer or on a Bruker AMX-400 spectrometer
1
operating at 360.13 (or 400.14) MHz for H and 90.6
Table 3
Fluorescence data on the complexes formed between Eu3+ and cryptands 3 and 4; uexc=338 nm
Eu3+
Co+ (cryptand, 3) 1.2 mM, uex=338 nm
Eu3+
Fe (cryptand, 4) 1.5 mM, uex=338 nm
10 eq
50 eq
593 (0.98), 614 (1.91)
555 (0.66), 560 (0.56), 593 (1.77), 616 (1.86)
–
–
50 eq
579.5 (0.87), 593 (1.61), 615 (2.75)

C.D. Hall, N. Djedo6ic / Journal of Organometallic Chemistry 648 (2002) 8–13
13
Fluorescence spectra were measured with a Perkin–
Elmer Luminescence Spectrometer LS50B and analyzed
with FL WinLab Version 2.01 Software using either the
default scan or time drive methods. The excitation and
emission slits were 10 and 2.5 nm, respectively. Fluores-
cence spectra (excited at 338 nm) of the ligands (1.2–
1.5 mM) and complexes with Eu3+ were measured in
MeCN at room temperature.
minimum of CH3CN. Upon dropwise addition to ether
(20 ml), a yellow solid precipitated (53.0 mg, 26%).
HRMS-FAB for C32H30O4N4Co, Mcalcd 593.1621, Mobs
593.1549. Anal. Found: C, 52.27; H, 4.11; N, 7.54.
Calc.: C, 52.05; H, 4.09; N, 7.59%.
Acknowledgements
3.1. Preparation of calcium trifluoromethanesulfonate
We would like to thank the EPSRC for a research
studentship (to N.D.) and the Intercollegiate Research
Service of the University of London for NMR (Mrs. J.
Hawkes and Mr. J. Cobb (KCL) and mass spectra (Dr.
K. Welham, School of Pharmacy).
Trifluoromethanesulfonic acid (0.89 g, 5.92 mmol)
dissolved in dry MeCN (1 ml) was added to a suspen-
sion of CaCO3 (0.3 g, 2.96 mmol) in dry MeCN (4 ml)
with stirring. After 15 min, insoluble starting material
was filtered off and the filtrate was evaporated to
dryness. The product was dried under vacuum (130 °C,
6.0–7.0×10−2 mbar) for 2 h to give a white crystalline
solid (0.71 g, 64%).
References
[1] P.D. Beer, P.A. Gale, G.Z. Chen, Coord. Chem. Rev. 185–186
(1999) 3.
[2] G.W. Gokel, A.E. Kaifer, S. Mendoza, Comp. Supramolec.
Chem. 1 (1996) 701.
3.2. Preparation of cryptand 3
A creased, round-bottomed, three-necked flask was
charged under N2 with a mixture of dry CH2Cl2 (250
ml), dry CH3CN (250 ml), and Et3N (0.40 ml, 2.84
mmol). The diazamacrocycle [13] (0.10 g, 0.28 mmol)
was dissolved in CH2Cl2 (250 ml) and cobaltocenium
chloride diacid chloride [11] (0.10 g, 0.30 mmol) was
dissolved in CH3CN (250 ml). The two solutions were
added to the stirred contents of the flask simultaneously
over 6 h using electrically powered syringes and the
mixture was left stirring for 15 h. The suspension was
filtered and the filtrate concentrated in vacuo. The
residue was redissolved in CH2Cl2 (50 ml), washed with
0.1 M aqueous NaOH and 10% aqueous NH+4 PF6−,
before drying over MgSO4. The solvent was removed
under reduced pressure and the yellow residue was
purified by flash chromatography on silica (CH2Cl2:
CH3OH) to give a yellow oil that was dissolved in a
[3] G.W. Gokel, J.C. Lockhart, Comp. Supramolec. Chem. 1 (1996)
605.
[4] T.E. Edmonds (Ed.), Chemical Sensors, Blackie, Glasgow, Lon-
don, 1988, p. 191 (Chapter 8).
[5] G.W. Gokel inc. J.F. Stoddart (Ed.), Crown Ethers and
Cryptands, R.S.C. Monograph, 1991.
[6] C.D. Hall, T.K.U. Truong, J. Organomet. Chem. 519 (1996) 185.
[7] J.H.R. Tucker, Ph.D. Thesis, University of London, 1993.
[8] A.P. Bell, P.J. Hammond, C.D. Hall, J. Chem. Soc. Perkin
Trans. 1 (1983) 707.
[9] C.D. Hall, T.K.U. Truong, S.C. Nyburg, J. Organomet. Chem.
547 (1997) 281.
[10] C.D. Hall, S.Y.F. Chu, J. Organomet. Chem. 498 (1995) 185.
[11] (a) C.D. Hall, J.H.R. Tucker, N.W. Sharpe, Organometallics 10
(1991) 1727;
(b) C.D. Hall, J.H.R. Tucker, S.Y.F. Chu, A.W. Parkins, S.C.
Nyburg, J. Chem. Soc. Chem. Commun. (1993) 1505.
[12] C.D. Hall, I.P. Danks, S.C. Nyburg, A.W. Parkins, N.W.
Sharpe, Organometallics 9 (1990) 1602.
[13] J.E. Sheats, M.D. Rausch, J. Org. Chem. 35 (1970) 3245.