Lanthanide triple-stranded helical complexes with a substituted 2,6-pyridinedicarboxylate

Article abstract of DOI:10.1039/b104448f

A 4-aminophenylethynyl substituent has been grafted into the 4-position of the pyridine ring in dipicolinic acid to yield ligand L⁴ the pK_a's of which are 8.77(5), 3.54(3), 2.58(1) and 1.01(3). Ligand L⁴ reacts with trival

Full text of DOI:10.1039/b104448f

View Article Online / Journal Homepage / Table of Contents for this issue
Lanthanide triple-stranded helical complexes with a substituted
2,6-pyridinedicarboxylate†
Carlos Platas-Iglesias,a,b Claude Piguet,c Nicolas Andréa and Jean-Claude G. Bünzli*a
a Institute of Inorganic and Analytical Chemistry, BCH, University of Lausanne,
CH-1015 Lausanne, Switzerland
b Departamento de Química Fundamental, Universidade da Coruña, Campus de A Zapateira s/n,
15071 A Coruña, Spain
c Department of Inorganic, Analytical and Applied Chemistry, University of Geneva,
30 quai E. Ansermet, CH-1211 Geneva 4, Switzerland
Received 21st May 2001, Accepted 3rd August 2001
First published as an Advance Article on the web 1st October 2001
A 4-aminophenylethynyl substituent has been grafted into the 4-position of the pyridine ring in dipicolinic acid to
yield ligand L4 the pKa’s of which are 8.77(5), 3.54(3), 2.58(1) and 1.01(3). Ligand L4 reacts with trivalent lanthanide
ions to give stable triple-stranded helical complexes in water (logβ13 = 13.7(1) for Eu). The separation of the contact
and pseudocontact contributions to the NMR paramagnetic shifts according to the crystal field independent method
points to a single axial structure for [Ln(L4 Ϫ 2H)3]3Ϫ along the series Nd–Yb. The solution structure generated by
using the dipolar shifts compares well with that reported for the EuIII tris(dipicolinate), with a co-ordination
polyhedron which can be described as a distorted tricapped trigonal prism. This is confirmed by a photophysical
study of the EuIII complex both in solution and in the solid state. The triplet state of co-ordinated L4 lies at low
energy (around 18 000 cmϪ1 in water) resulting in a relatively poor sensitisation of EuIII (quantum yield ca. 0.1% in
methanol, lower in water) but in a sizable energy transfer to YbIII.
Introduction
Lanthanide co-ordination compounds1 continue to receive
much attention owing to their important applications as con-
trast agents for NMR imaging,2 catalysts for RNA hydrolysis,3
active agents in cancer radiotherapy,4 or as luminescence stains
in bioanalyses and medical diagnostic.5 In recent years, we have
developed a research program based on the induced fit concept6
to prepare mono- and di-metallic functional edifices.7 Our
strategy involves the use of aromatic tridentate units derived
either from dipicolinic acid (2,6-pyridinecarboxylic acid)8 or
from bis(benzimidazole)pyridine (Scheme 1). The latter binding
units have been successfully introduced into homotopic and
heterotopic ligands which strictly self-assemble with LnIII ions
to produce triple-stranded cationic [Ln2(L)3]6ϩ 9,10 or neutral
[Ln2(LC Ϫ 2H)3]11 f–f helicates as well as triple-stranded d–f
heterodimetallic podates [LnM(LЈ)]5ϩ (M = Zn, Fe,12 Co13). In a
recent paper we have developed a synthetic strategy for grafting
an halogenide substituent onto the pyridine moiety of ligand L
to obtain LE, which can be used as a synthon for additional
substitution of the pyridine with more elaborate groups.10
Luminescent probes for bioanalytical analyses require
assemblies that are stable under physiological conditions,
which bear efficient chromophoric groups and which can easily
couple with biological molecules. As an alternative to the above
strategy, we test here the possibility of grafting adequate sub-
stituents in the 4-position of the dipicolinic acid framework,
since this tridentate aromatic unit forms stable and strongly
luminescent triple-stranded helical [Ln(L1 Ϫ 2H)3]3Ϫ complexes
in water14 or [Ln(L2,3)3]3ϩ edifices in acetonitrile.15 We have also
† Electronic supplementary information (ESI) available: analytical, IR
and luminescence data, contact and dipolar shifts, calculated Cartesian
co-ordinate for the EuIII complex, plots of chemical shifts for protons
H8, H9 versus the chemical shift of proton H2. See http://www.rsc.org/
suppdata/dt/b1/b104448f/
Scheme 1
3084
J. Chem. Soc., Dalton Trans., 2001, 3084–3091
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b104448f

View Article Online
recently demonstrated that the introduction of a bulky chiral
substituent in the R2 position of L2 does not prevent helical
wrapping around the metal ions and leads to the formation
of stable complexes in acetonitrile.16 In another study, Wensel
et al.17 have examined the sensitisation of TbIII in water by a
series of 4-substituted analogues of dipicolinic acid bearing
simple groups (R = NH2, OH, NHAc, Cl, H or Br). In this
work, we introduce a 4-aminophenylethynyl substituent in the
4-position of the pyridine in L1 to yield L4. It is expected that
electronic delocalisation will induce a relatively low lying triplet
state, thus providing energy transfer paths to EuIII and YbIII.
We also investigate the solution structure of the resulting 1 : 3
complexes making use of the separation of contact and pseudo-
contact contributions to the NMR paramagnetic shifts.
cedure to that previously reported by Takalo and co-workers,24
from 4-aminophenylacetylene25 and 4-bromopyridine-2,6-di-
methylcarboxylate17 instead of the corresponding iodine-
derivative: 1.124 g of 4-bromopyridine-2,6-dimethylcarboxylate
(4.1 mmol), 0.0581 g of bis(triphenylphosphine)palladium()
chloride (0.082 mmol) and 0.0312 g of copper() iodide (0.164
mmol) were suspended in 16 cm3 of Et3N and the mixture was
stirred under an inert atmosphere (N2) for 15 min. A 0.480 g
amount of 4-aminophenylacetylene (4.1 mmol) was added
and the mixture was heated under N2 for 4 h at 40 ЊC and then
evaporated to dryness. The solid residue was dissolved in
CH2Cl2 (200 cm3) and the resultant solution washed with
water (30 cm3). The organic phase was dried over Na2SO4, and
evaporated to dryness. Methanol (20 cm3) was added to the
residue, a yellow-brown crystalline product appeared which was
filtered off under vacuum, washed with MeOH (20 cm3) and
dried under vacuum to get 1.081 g of the desired compound
(yield 85%). 1H-NMR in (CD3)2SO: δ 8.16 (2H, s), 7.33 (2H, t,
3J = 8.94 Hz), 6.58 (2H, t, 3J = 8.93), 5.86 (2H, b), 3.93 (6H, s).
Experimental
Solvents and starting materials
Triethylamine was distilled from CaH2. Other products were
purchased from Fluka AG (Buchs, Switzerland) and used
without further purification, unless otherwise stated. The Na3-
[Gd(L1 Ϫ 2H)3] complex was prepared by a similar procedure to
that previously described.14b
4-(4-Aminophenylethynyl)pyridine-2,6-dicarboxylic acid di-
chlorohydrate, L4ؒ2HCl. A 1.081 g (3.48 mmol) amount of 4-
(4-aminophenylethynyl)pyridine-2,6-dimethylcarboxylate was
refluxed in 50 cm3 of a KOH solution 0.1 M in EtOH–H2O
(1 : 1). After completion of the reaction (12 h) the solution was
concentrated under vacuum to remove the ethanol and the
resulting brown aqueous solution was acidified to pH ≈ 1 with
25% aqueous HCl. The brown precipitate formed was filtered
off, washed with water (5 cm3) and dried under vacuum (10Ϫ2
Torr) at 75 ЊC for 12 h to get 0.854 g of the desired product as
a pale brown solid (yield 69%) [Found: C, 51.43; H, 3.15; N,
7.92; calc. for C15H10N2O4ؒ2HCl: C, 50.73; H, 3.41; N, 7.89%].
1H-NMR in (CD3)2SO: δ 8.02 (2H, s), 7.36 (2H, d, 3J = 8.9 Hz),
Spectroscopic and analytical measurements
Elemental analyses were carried out on a Carlo Erba EA 1108
elemental analyser. Electronic spectra in the UV-vis range were
recorded at 20 ЊC with a Perkin-Elmer Lambda 7 spectrometer
using 1.0 and 0.1 cm quartz cells. In a typical spectrophoto-
metric pH titration, 25 cm3 of a 1 × 10Ϫ5 M solution of L4 or
of a 1.5 × 10Ϫ4 M solution of Cs3[Eu(L4 Ϫ 2H)3] in water con-
taining 0.05 M KClO4 was prepared, and the starting pH
(measured with a Metrohm 6.3013.210 glass electrode) was
adjusted with solid Cs2CO3. The pH was then progressively
decreased with 1 M HCl, the equilibrium being reached almost
immediately. The spectrophotometric titration of L4 with
Eu(ClO4)3ؒxH2O was performed on a Perkin-Elmer Lambda
900 spectrometer: 25 cm3 of a 2.53 × 10Ϫ5 M solution of L4
buffered at pH 7.7 with 0.1 M tris/trisؒHCl [tris = tris(hydroxy-
methyl)aminomethane] was prepared, and 0.05 cm3 aliquots of
a 2.68 × 10Ϫ4 M solution of Eu(ClO4)3ؒxH2O were successively
added. For both titrations, plots of the molar absorption co-
efficient as a function of the pH or M/L ratio as well as factor
analysis18 were used to determine a suitable model, and the data
were analysed with the SPECFIT program.19 Reflectance spec-
tra were recorded as finely ground powders dispersed in MgO
(5%) with MgO as the reference on a Perkin-Elmer Lambda 900
spectrometer equipped with a Labsphere PELA-1000 inte-
gration sphere. IR spectra were obtained from KBr pellets with
3
6.65 (2H, d, J = 8.9). ESI-MS (H2O): m/z 283 [M ϩ H]ϩ.
IR ν(cmϪ1, KBr): 2200 (C᎐C), 1719 (C᎐O), 1116 (C–O).
᎐
᎐
᎐
Preparation of the Cs3[Ln(L4 ؊ 2H)3]ؒxH2O complexes
The perchlorate salts Ln(ClO4)3ؒxH2O (x = 6–8) were prepared
from the corresponding oxides (Rhône-Poulenc, 99.99%)
according to the literature method.26
CAUTION: perchlorate salts combined with organic ligands
are potentially explosive and should be handled in small
quantity and with the necessary precautions.27
A 0.030 g amount of L4ؒ2HCl (84.5 µmol) was suspended in
methanol (10 cm3) and Cs2CO3 (0.0358 g, 337.9 µmol) were
added. The resultant solution was stirred and heated to reflux
for 10 min; the corresponding lanthanide perchlorate salt
(28.15 µmol) dissolved in 10 cm3 of the same solvent was added
dropwise to the former solution. The resultant yellow solution
was stirred and heated to reflux until complete reaction of the
carbonate, filtered while hot and evaporated slowly at room
temperature. The yellow crystals formed were filtered and dried
under vacuum to give Cs3[Ln(L4 Ϫ 2H)3]ؒnH2O (n = 2–3, Ln =
Eu, Gd, Tb) in 70–80% yield. The [Ln(L4 Ϫ 2H)3]3Ϫ complexes
used for NMR measurements (Ln = Nd, Tb, Dy, Ho, Er, Tm,
Lu, Y) were prepared by the same procedure, but were not
crystallised.
1
a Mattson α-Centauri FT-IR spectrometer. H and 13C NMR
spectra were recorded at 25 ЊC on Bruker AM-360 or Bruker
AVANCE 400-DRX spectrometers, and the pD (= pH ϩ 0.420)
was adjusted with D2SO4 (Fluka) or NaOD (40% in D2O,
Merck). Chemical shifts are reported in parts per million with
respect to SiMe4 (referenced to internal HOD at 4.75 ppm).
The experimental procedures for high resolution, laser excited
luminescence measurements have been published previously.21
Quantum yields were measured relative to quinine sulfate in
0.05 M H2SO4 according to a procedure previously described22
(n = 1.338, absolute quantum yield: 0.546).23 The refractive
indices used were 1.333 and 1.329 for solutions in H2O and
MeOH, respectively. Ligand excitation and emission spectra
Results and discussion
Synthesis and properties of the ligand and its lanthanide
complexes
were recorded on
a Perkin-Elmer LS-50B spectrometer
equipped for low temperature (77 K) measurements.
4-(4-Aminophenylethynyl)pyridine-2,6-dicarboxylic acid (L4)
was obtained in good overall yield from commercially available
products by a four-step procedure involving a Sonogashira28
type coupling reaction as the key step for the substitution of the
pyridine 4-position (Scheme 2). At pH = 10.74, [L4 Ϫ 2H]2Ϫ
Preparation of the ligand
4-(4-Aminophenylethynyl)pyridine-2,6-dimethylcarboxylate.
The compound was prepared by a somewhat different pro-
J. Chem. Soc., Dalton Trans., 2001, 3084–3091
3085

View Article Online
Scheme 2
absorption spectra of a 1.5 × 10Ϫ4 M aqueous solution of
[Eu(L4 Ϫ 2H)3]3Ϫ remain unchanged in the pH range 5.00 < pH
< 10.4, while dissociation of the complex occurs at pH < 5.0.
Upon complexation, the π π* transition of L4 appearing at
higher energy undergoes a red shift of ca. 1750 cmϪ1, thus
enabling the monitoring of the complex formation.
Spectrophotometric titration of L4 with Eu(ClO4)3ؒxH2O
was performed in aqueous solution buffered at pH 7.7. Plots
of the molar absorption coefficient versus the ligand/metal ratio
at different wavelengths show a break for the 3 : 1 complex,
the spectrum remaining then unchanged until the ligand/metal
ratio reaches 1.35 : 1. This result was confirmed by factor
analysis and the data could be satisfactorily fitted to equi-
librium (1) allowing an estimation of the overall stability con-
Eu
3 (L4 Ϫ 2H)2Ϫ ϩ Eu3ϩ
[Eu(L4 Ϫ 2H)3]3Ϫ log(β13
)
(1)
stant: log(β13Eu) = 13.7(1). Protonation of the amine group of
L4 was not taken into account for the sake of simplicity.
This result should be taken with care, since precipitation
occurs for ligand/metal ratios < 1.35 : 1, preventing a complete
analysis allowing for the formation of the 1 : 1 and 1 : 2 species.
Nevertheless, the log(β13Eu) value points to an important
decrease in the stability of the [Ln(L4 Ϫ 2H)3]3Ϫ complexes
compared to [Ln(L1 Ϫ 2H)3]3Ϫ, for which log(β13Ln) is in the
range 18–22.15a This can be tentatively assigned to a larger
electron delocalisation onto the extended π system of L4, but
partial protonation of the terminal amino groups leading to
reduced electrostatic Ln–ligand interactions in [Ln(L4 Ϫ 2H ϩ
nH)3](3 Ϫ n)Ϫ cannot be excluded at pH = 7.7.
Fig. 1 Distribution diagram for L4.
presents two main absorption bands at 29 918 and 38 760 cmϪ1
,
assigned to π π* transitions mainly located on the 4-amino-
phenylethynyl and pyridine moieties, respectively. A spectro-
photometric titration of a 10Ϫ5 M aqueous solution of [L4 Ϫ
2H]2Ϫ with hydrochloric acid (starting pH = 10.76) revealed four
different equivalence points corresponding to pKa1 = 8.77
0.05, pKa2 = 3.54 0.03, pKa3 = 2.58 0.01 and pKa4 = 1.01
0.03. The distribution curves of the different species are shown
in Fig. 1. The highest protonation constant corresponds to the
protonation of the amine group, while the second represents the
protonation of the pyridine nitrogen, the latter value being
ca. 10 times smaller than for dipicolinic acid (pKa = 4.63,29
4.7630), probably as a consequence of the reduced negative
charge in [L4 Ϫ H]Ϫ compared to [L1 Ϫ 2H]2Ϫ. Finally, the
last two protonation steps with a ∆pKa of 1.57 correspond to
the protonation of the carboxylic functions. As the expected
statistical separation for two non-interacting carboxylic groups
belonging to the same molecule is ∆pKa = 0.6,31 the larger
∆pKa value observed for L4 points to the two carboxylic groups
interacting through intramolecular hydrogen bond(s) in the
[L4 ϩ H]ϩ species.
Solution structure of [Ln(L4 ؊ 2H)3]3؊ complexes
The structure of the [Ln(L4 Ϫ 2H)3]3Ϫ complexes was studied
1
in D2O by H NMR spectroscopy, since the low solubility of
the complexes prevents recording the 13C spectra. The spectra
of the Y and Lu diamagnetic complexes (Table 2) show the
expected simple pattern with three signals corresponding to the
three non-equivalent aromatic protons of the three equivalent
ligand strands. The signal due to the H2 protons shifts to higher
field by 0.11 ppm upon complexation, while the signals of the
H8 and H9 protons are less affected. The 1H NMR shifts of the
paramagnetic complexes are also listed in Table 2, as measured
at 6 < pD < 8. A detailed study of a 10Ϫ3 M aqueous solution of
Reaction of stoichiometric amounts of L4ؒ2HCl, Cs2CO3
and the corresponding hydrated lanthanide perchlorate (Ln =
Eu, Gd, Yb) gives yellow-orange solutions in methanol from
which the Cs3[Ln(L4 Ϫ 2H)3]ؒnH2O complexes can be isolated
in 70–80% yield (Table S1, ESI). These complexes present a
number of identifying IR bands including (see Table S1,
ESI): (i) intense carbonyl symmetric and asymmetric stretching
vibrations at 1590 and 1616 cmϪ1, respectively, (ii) an intense
1
[Eu(L4 Ϫ 2H)3]3Ϫ vs. the pD shows the H shifts being com-
pletely pD-independent between 5.7 and 10.3, suggesting that
the EuIII complex presents a single solution structure in this pD
range, in agreement with the UV-visible data (vide supra).
Finer structural information can be gained from the
separation of the contact (δcij) and pseudocontact contributions
C᎐C stretching vibration at ca. 2200 cmϪ1, and (iii) unco-
᎐
᎐
pc
ij
para
ij
32
ordinated water molecules at 3400 cmϪ1. Moreover, the absence
(δ ) to the isotropic paramagnetic shift (δ
)
of a nucleus i
induced by a lanthanide ion j (LIS):
of any absorption band arising from the perchlorate group
unambiguously proves the displacement of the anion during
δpijara = δiejxp Ϫ δdi ia = δicj ϩ δipjc = Fi 〈Sz〉j ϩ GiCj
(2)
the formation of the complexes. At pH 10.4 the [Ln(L4
Ϫ
2H)3]3Ϫ complexes present two main absorption bands (Table
1), which are shifted to lower energies compared to those of
[L4 Ϫ 2H]2Ϫ. In complete contrast with the free ligand, the
where Fi is the hyperfine coupling constant of nucleus i,
which governs the contact interaction between that nucleus and
3086
J. Chem. Soc., Dalton Trans., 2001, 3084–3091

View Article Online
Table 1 Ligand-centred absorption and emission properties of L4 and [Ln(L4 Ϫ 2H)3]3Ϫ
Compound
L4
E(π π*)a/cmϪ1
Solvent
E(π π*)b/cmϪ1
E(1ππ*)d/cmϪ1
E(3ππ*)e/cmϪ1
τ(3ππ*)e/ms
29 233
36 605
H2Oc
29 918 (4.16)
38 760 (4.04)
27 585 (4.15)
37 678 (4.03)
28 169 (4.68)
38 168 (4.01)
46 125 (4.70)
27 585 (4.08)
37 678 (4.00)
47 092 (4.20)
28 558 (4.67)
38 121 (4.59)
46 685 (4.77)
27 777 (4.64)
37 594 (4.51)
46 729 (4.72)
22 844
19 176
20 175
MeOH
H2Oc
18 727
Eu
28 203
35 540
f
f
MeOH
H2Oc
Gd
27 864
35 540
18 132
19 656
17 555
18 198
1.54 0.03
1.86 0.09
MeOH
a Reflectance spectra recorded at 295 K. b Electronic spectral data at 295 K; energies are given for the maximum of the band envelope in cmϪ1; log ε is
given within parentheses. c Buffered solution (pH = 8.0). d From luminescence data in solution at 295 K. e Luminescence data and lifetimes in frozen
solutions at 77 K. f 3ππ* Luminescence quenched by transfer to the lanthanide ion.
Table 2 1H-NMR shifts (ppm) for L4 (pD = 12.0) and for the [Ln(L4 Ϫ 2H)3]3Ϫ complexes in D2O (7.4 < pD < 8.4) at 298 K, and relevant data for the
analysis of the paramagnetic shifts
Compound
H2
H8
7.41
7.44
7.88
H9
6.77
6.78
6.98
AFj c
AFj d
AFj e
(L4 Ϫ 2H)2Ϫ
Y
Nd
Eu
Tb
Dy
Ho
Er
Tm
Yb
Lu
7.95
8.06
10.49
3.89
0.2800
0.0519
0.1420
0.0300
0.0189
0.0503
0.0881
0.0453
0.0065
0.0067
0.0055
0.0056
0.0153
0.0154
0.0151
0.0150
0.1368
0.1364
0.1388
0.1387
0.1404
0.1407
0.1407
0.1400
6.75
6.48
41.52
44.92
27.34
Ϫ2.38
Ϫ13.89
0.04
15.47
16.09
13.27
5.41
3.04
5.64
11.38
11.69
10.29
5.60
4.37
5.79
8.06
7.39
6.73
Nd to Dy
Fi
H9–H2
H9–H8
H8–H2
0.29(7)
0.44(3)
0.1009
0.03(1)
0.096(5)
0.0630
0.008(5)
0.052(3)
0.0580
Gi
AFi c
a
Rik
0.118
0.542
0.218
Ho to Yb
H9–H2
H9–H8
H8–H2
Fi
Ϫ0.16(4)
0.40(3)
0.0631
Ϫ0.089(8)
0.092(3)
0.0590
Ϫ0.057(8)
0.0512(3)
0.0968
Gi
AFi c
a
Rik
Nd to Yb
0.128
0.557
0.230
H9–H2
0.122(4)
Ϫ0.025(6)
H9–H8
0.552(6)
Ϫ0.007(2)
H8–H2
0.221(6)
Ϫ0.032(9)
b
Rik
(Fi Ϫ RikFk)b
a Rik = Gi /Gk, with Gi and Gk obtained according to eqn. (4) and (5). b According to eqn. (7). c Reilley’s analysis, eqn. (6). d Calculated shifts estimated
by the SHIFT ANALYSIS program42 with assignment shown (see text). e Calculated shifts estimated by the SHIFT ANALYSIS program42 with
inverted assignment for H8 and H9 (see text).
the Lnj ion and Gi is the geometric factor of nucleus i con-
taining the structural information, 〈Sz〉j and Cj are, respectively,
the spin expectation value and the magnetic constant (Bleaney
factor) of the paramagnetic lanthanide. The diamagnetic con-
tribution δdi ia is obtained by measuring chemical shifts for iso-
structural diamagnetic complexes: [Y(L4 Ϫ 2H)3]3Ϫ was used for
the earlier (Nd–Dy) and [Lu(L4 Ϫ 2H)3]3Ϫ for the later (Ho–Yb)
members of the lanthanide series. For a complex with effective
axial magnetic symmetry,33
the ligand field axis. A straightforward separation of the con-
tact term Fi and the pseudo-contact term Gi for each nucleus i
in the axial lanthanide complexes [Ln(L4 Ϫ 2H)3]3Ϫ can be per-
formed by rearranging eqn. (2) into the linear forms given by
eqn. (4) and (5):
(4)
(5)
(3)
Assuming that 〈Sz〉j and Cj values are the same for the com-
plexes and the free ions,34 for which they are tabulated,35,36 plots
of δipjara/〈Sz〉j against Cj /〈Sz〉j and of δipjara/Cj against 〈Sz〉j /Cj should
be linear with a slope equal to Gi and Fi, respectively, if the
complexes are isostructural and possess comparable crystal
field parameters. The pseudocontact and contact contributions
where A02〈r2〉 is the second-order axial crystal field parameter,
which is a measure of the strength of interaction between a
given lanthanide ion and the ligand donor atoms, and ri and θi
are the internal axial co-ordinates of nucleus i with respect to
J. Chem. Soc., Dalton Trans., 2001, 3084–3091
3087

View Article Online
The three plots according to eqn. (7) involving the H2, H8 and
H9 protons display a single straight line (Nd–Yb, Fig. 2b), and
yield Rik values in excellent agreement with those calculated
from eqn. (4) and (5) (Table 2), confirming a single structure
along the series (Nd–Yb) and also that a change in the hyper-
fine coupling constant is responsible for the breaks found in the
plots according to eqn. (4) and (5). This behaviour contrasts
with that of the [Ln(L2)3]3ϩ complexes, which show two dif-
ferent structural series for Ln = Ce–Tb and Ln = Er–Yb.39
As eqn. (7) takes into account the paramagnetic shift for two
nuclei i and k, two new agreement factors characteristics of the
fitting process and related to a given proton (eqn. (8)) or a given
lanthanide ion (eqn. (9)) can be calculated:
(8)
(9)
cal
Where δ f(δekxjp) is defined as the paramagnetic shift for
ij
nucleus i calculated from eqn. (7) by using the experimental
paramagnetic shift for nucleus k. The extracted agreement
factors (Table S3, ESI) are globally satisfactory and are in line
with the AFi’s and AFj’s obtained for the plots according to
cal
exp
eqn. (4) and (5). However, we notice that the use of δ f(δ
)
ij
9j
for H9 and, in a less drastic proportion, for H8 as references
produce poor predictions according to eqn. (7) for the weakly
paramagnetic ions NdIII and EuIII. This may be explained
by the fact that these remote protons undergo only small
para
Fig. 2 Plots of: (a): δ /〈Sz〉j vs. Cj /〈Sz〉j for H9 according to eqn. (4)
ij
paramagnetic induced shifts and, therefore, the division by
para
and (b): δpijara/〈S 〉 vs. δ /〈Sz〉j (i = H9, k = H8) according to eqn. (7) for
exp
small Σ(δ )2 quantities leads to large AFi,k’s and AFj,k’s for
kj
z
j
ij
[Ln(L4 Ϫ 2H)3]3Ϫ 10Ϫ3 M in D2O at 298 K.
these ions.
We have also used the dipolar shifts to generate the structure
of the [Ln(L4 Ϫ 2H)3]3Ϫ complexes in aqueous solution, which
requires an initial estimate of this structure. It is well known
that [Ln(L1 Ϫ 2H)3]3Ϫ complexes present a somewhat distorted
D3 symmetry in the solid state with a co-ordination polyhedron
around the lanthanide ion close to a tricapped trigonal
prism.14b,c On the other hand, circularly polarised luminescence
studies are consistent with a D3 symmetry in solution, the
complexes existing as a racemic mixture of Λ and ∆ enantio-
meric structures.40 Since luminescence data demonstrate that
the [Eu(L4 Ϫ 2H)3]3Ϫ complex also displays D3 symmetry in
solution (see below), we have taken the X-ray structure of the
[Eu(L1 Ϫ 2H)3]3Ϫ complex as a reasonable starting model for
the structure of [Eu(L4 Ϫ 2H)3]3Ϫ. The 4-aminophenylethynyl
substituent was then introduced using the molecular mechanics
program Cerius-241 in the following way: the co-ordinates
of the donor atoms were fixed and the molecule let to relax,
yielding a first set of Cartesian co-ordinates. The C3 axis of
symmetry, taken as the z-axis, was assumed to go through the
metal ion and one of the three triangular faces of the trigonal
prism formed by three carboxylate O atoms. One of the Ln–N
vectors was set as the y axis. Finally, the Ln(L4 Ϫ 2H) unit with
the pyridine nitrogen atom situated on the y axis was used to
estimate the idealised Ln–donor distances with the SHIFT
ANALYSIS program:42 the principal magnetic axis of sym-
metry was taken as the z axis, the Ln ion was moved along the y
axis and plots of the agreement factor AFj (eqn. (6)) vs. the
Ln–N bond distance revealed minima that best fit the LIS data
(Fig. 3). The obtained idealised Ln–N and Ln–O distances
were 2.57 and 2.46 Å, respectively, in good agreement with the
averaged distances found for [Eu(L1 Ϫ 2H)3]3Ϫ in the solid state
(2.54 and 2.45 Å, respectively).14c In a last step, the resulting
Cartesian co-ordinates obtained from the optimisation of the
Ln–donor distances (Table S4, ESI) were used, after trans-
formation into cylindrical (r, θ, z) co-ordinates and assuming
trigonal symmetry, to generate the best structural model for the
to the observed LIS separated by using eqn. (4) and (5) are
reported in Table 2 along with the agreement factors AFi and
AFj (eqn. (6); see also Table S2, ESI). The resulting plots fall
(6)
into two groups (Ln = Nd–Dy and Ln = Ho–Yb) with a break
near the middle of the series as illustrated in Fig. 2a for H9. In
principle, such a situation may point to the [Ln(L4 Ϫ 2H)3]3Ϫ
helicates being not isostructural. However, deviation from
linearity also occurs because of variation in the crystal field
parameter37 or in both this parameter and the hyperfine
coupling constant Fi.38,39 In our case, since two straight parallel
lines are observed we can infer that: (i) the complexes are most
probably isostructural along the lanthanide series, (ii) the
crystal field parameter remains invariant along the lanthanide
series, but (iii) a change occurs in the contact terms along the
lanthanide series and is responsible for the break. The agree-
ment factors calculated for both series, 0.058 < AFi < 0.101
(Nd–Dy) and 0.059 < AFi < 0.097 (Ho–Yb), are comparable to
those found for [Ln(L1 Ϫ 2H)3]3Ϫ 34 and 4f–4f11 helicates. The
contact contributions are relatively small for all the protons,
Fi reaching a significant value only for H2, which points to a
sizeable spin density delocalisation onto the pyridine ring.
Recently, Platas et al.38 proposed a crystal-field independent
method, which allows the rationalisation of the origin of some
of the breaks found in the plots according to eqn. (4) and (5), by
simultaneously solving eqn. (2) for two different nuclei i and k:
(7)
3088
J. Chem. Soc., Dalton Trans., 2001, 3084–3091

View Article Online
Fig. 5 Phosphorescence spectra at 77 K for: (a) L4 in frozen methanol
solution; (b) [Gd(L4 Ϫ 2H)3]3Ϫ in frozen methanol solution; (c) [Gd-
(L4 Ϫ 2H)3]3Ϫ in frozen water solution; (d) [Eu(L4 Ϫ 2H)3]3Ϫ in frozen
methanol solution.
Fig. 3 Plot of the agreement factor AFj vs. the Ln–N bond distance for
[Eu(L4 Ϫ 2H)3]3Ϫ in D2O.
(15.6Њ),14c despite the reduced number of experimental data at
hand to refine the solution structure.
Finally, we have checked the assignment for protons H8
and H9 by using the SHIFT ANALYSIS program42 and the
above-calculated geometry to test an interchange of attribution
between these two protons for each LnIII ion. The calculated
pseudocontact contributions to the LIS match those extracted
from Reilley’s analysis when the attribution of Table 2 is taken
into consideration, which is not the case for the inverted
assignment. This is reflected in the corresponding AFi’s
calculated according to eqn. (6) and reported in Table 2. It
is noteworthy that plots of the chemical shifts of H8 and H9
assigned according to Table 2 versus the chemical shift of H2,
the assignment of which is unambiguous, are linear, reflecting
both the correctness of the H8 and H9 signal assignment and the
isostructurality within the series (Fig. S1, ESI).
Photophysical properties
The emission spectrum of L4 in methanolic solution (295 K)
under excitation at 25 253 cmϪ1 presents a single weak emission
band (Table 1) the intensity of which quickly diminishes when a
short delay (0.1 ms) is enforced and this band is therefore
1
attributed as arising from the ππ* state. In aqueous solution
1
the ππ* emission appears at higher energy and is nearly pH-
independent in the range 5.6–10.9, in complete contrast with
the absorption spectra. The emission spectrum recorded in
frozen methanolic solution (77 K, delay 0.10 ms) presents a
second band with a maximum at 18 730 cmϪ1 which is assigned
as arising from to the 3ππ* state (Fig. 5). Upon complexation to
the GdIII ion the ligand-based triplet state emission under-
goes a red shift of 529 cmϪ1 (MeOH) or 643 cmϪ1 (H2O). The
introduction of a 4-aminophenylethynyl substituent affects
Fig. 4 Perspective view of one of the two helical forms of the [Ln(L4
1
Ϫ 2H)3]3Ϫ complex in D2O solution as determined from the H-NMR
data.
complexes in solution (Fig. 4). The cylindrical co-ordinates were
transformed back into Cartesian co-ordinates (Table S4, ESI),
and the SHIFT ANALYSIS program was used to calculate the
agreement factor between the experimental and calculated
shifts. The resulting agreement factor AFj = 0.00668 is excellent.
The generated co-ordination polyhedron is a tricapped trigonal
prism which can be considered as being comprised of three
tripods defining three parallel planes: the O atoms from the
carboxylates define the upper and lower tripods, while the
N atoms of the pyridine moieties delineate the capping tripod,
the lanthanide ion lying at the centre of the latter tripod. The
distance between the facial planes of the distal tripods is 3.22 Å,
comparable to that observed for [Eu(L1 Ϫ 2H)3]3Ϫ in the solid
state (3.33 Å). The angle between the projection vectors of the
two distal tripods on the plane formed by the capping tripod
3
drastically the energy of the ππ* state which is located ca.
6000 cmϪ1 below that of Na3[Gd(L1 Ϫ 2H)3]. The lifetime of the
3ππ* state emission, 1.54 ms (H2O), is also substantially affected
by the substitution of the pyridine being ten times shorter than
in Cs3[Gd(L1 Ϫ 2H)3].14b
In contrast, the ligand based singlet and triplet state emis-
sions completely disappear in the EuIII triple-helical edifice
while the typical line emission arising from the metal ion is
observed (Fig. 5), indicating sensitisation of the metal ion via a
3ππ*-to-metal energy transfer. In methanol, the metal-centred
emission is dominated by the 5D0 7F2 transition, as shown by
the integrated and corrected relative intensities of the 5D0 7FJ
transitions: 0.01, 1.00, 10.07, 0.09 and 1.55 for J = 0, 1, 2, 3 and
4, respectively. There is a single, very weak, 5D0 7F0 transition
located at 17 221 cmϪ1, in line with the presence of a single EuIII
co-ordination environment. The emission spectrum (cf. Table
S5, ESI) may be interpreted in terms of a pseudo-D3 symmetry,
and the lanthanide ion amounts to 15.8Њ for the [Eu(L4
Ϫ
2H)3]3Ϫ complex, pointing to a certain degree of distortion
from the ideal geometry (angle 0Њ) but compares well with the
average value found in the solid state for [Eu(L1 Ϫ 2H)3]3Ϫ
5
in agreement with the NMR data: the D0 7F0 transition is
J. Chem. Soc., Dalton Trans., 2001, 3084–3091
3089

View Article Online
very weak (forbidden in D3), there are two components for the
The introduction of the 4-aminophenylethynyl substituent
drastically shifts the energy of the 3ππ* state of the ligand
toward lower energies, resulting in a poor sensitisation of EuIII,
but allows an efficient energy transfer onto YbIII which emits in
the near infrared. The 4-aminophenylethynyl substituent is
one of the groups commonly used for coupling luminescent
stains to biological materials47 so that the presently developed
synthetic technology will enable us to produce adequate
lanthanide-containing building blocks for the design of
luminescent probes, provided the problem of the (too) low lying
triplet state can be solved.
7
transition to F1 (allowed A1 A2 and A1 E transitions) and
7
two components for the transition to F2 (two allowed A1
E
transitions). The lifetime of the 5D0(Eu) state is short, amount-
ing to 0.177(4) ms, and the absolute quantum yield of the EuIII-
centred luminescence upon excitation through the ligand levels
is 0.09%. In water, the emission spectrum is of much weaker
intensity, the 5D0 lifetime dropping to ca. 0.02 ms (compared to
1.67 ms for [Eu(L1 Ϫ 2H)3]3Ϫ in the same conditions)43 and a
residual emission from the ligand 1ππ* is seen. It is known that
an efficient ligand-to-metal energy transfer requires a good
1ππ* 3ππ* intersystem crossing, which is maximised when the
energy difference between these states is close to 5000 cm .
Ϫ1 44 In
3
our case, ∆E(1ππ* Ϫ ππ*) amounts to 1458 and 577 cmϪ1 in
methanol and water, respectively, and is therefore too small.
Moreover, the energy difference between the 0-phonon band of
Acknowledgements
This work is supported by grants from the Swiss National
Science Foundation and by the Swiss Office for Education
and Science (COST action D18). We thank Veronique Foiret
for her help in recording high resolution luminescence spectra
and lifetimes and Dr M. Elhabiri and Mr M. Hollenstein for
preliminary work towards the synthesis of the ligand.
3
5
the ππ* emission and the D0 level, 337 (H2O) and 980 cmϪ1
(MeOH) as determined at 77 K, is also very small. This further
explains the poor sensitisation of the EuIII complex both in
5
water and in methanol, as well as the short lifetime of the D0
level, a back transfer process onto the ligand being possible.
A high resolution luminescence study has been performed at
low temperature (10 K) on a solid sample of Cs3[Eu(L4 Ϫ 2H)3]
to gain information on the chemical environment of the metal
References
1 J.-C. G. Bünzli, in Rare Earths, ed. R. Saez-Puche and P. Caro,
Editorial Complutense, Madrid, 1998, p. 223ff.
2 (a) R. B. Lauffer, in MRI Clinical Magnetic Resonance Imaging, ed.
R. R. Edelman M. B. Zlatkin and J. R. Hesselink, W. B. Saunders
Co., Philadelphia, 2nd ed., 1996, vol. 1, ch. 5; (b) P. Caravan,
J. J. Ellison, T. J. McMurry and R. B. Lauffer, Chem. Rev., 1999,
99, 2293.
3 (a) T. C. Bruice, A. Tsubouchi, R. O. Dempey and L. P. Olson,
J. Am. Chem. Soc., 1996, 118, 9867; (b) P. Hurst, B. K. Takasaki
and J. Chin, J. Am. Chem. Soc., 1996, 118, 9982.
4 (a) G. L. DeNardo, G. R. Mirik, L. A. Kroger, R. T. O’Donnel,
C. F. Meares and S. J. De Nardo, J. Nucl. Med., 1996, 37, 451;
(b) M. J. Heeg and S. S. Jurisson, Acc. Chem. Res., 1999, 32, 1053.
5 V. W. W. Yam and K. K. W. Lo, Coord. Chem. Rev., 1999, 184, 157.
6 D. E. Koshland, Jr., Angew. Chem., Int. Ed. Engl., 1994, 33,
2375.
7 C. Piguet and J.-C. G. Bünzli, Chem. Soc. Rev., 1999, 28, 347.
8 F. Renaud, C. Piguet, G. Bernardinelli, J.-C. G. Bünzli and
G. Hopfgartner, Chem. Eur. J., 1997, 3, 1646 and 1661.
9 C. Piguet, J.-C. G. Bünzli, G. Bernardinelli, G. Hopfgartner and
A. F. Williams, J. Am. Chem. Soc., 1993, 115, 8197; N. Martin,
J.-C. G. Bünzli, V. McKee, C. Piguet and G. Hopfgartner, Inorg.
Chem., 1998, 37, 577.
10 C. Platas, M. Elhabiri, M. Hollenstein, J.-C. G. Bünzli and C. Piguet,
J. Chem. Soc., Dalton Trans., 2000, 2031.
11 M. Elhabiri, R. Scopelliti, J.-C. G. Bünzli and C. Piguet, J. Am.
Chem. Soc., 1999, 121, 10747.
12 C. Edder, C. Piguet, G. Bernardinelli, J. Mareda, C. G. Bochet,
J.-C. G. Bünzli and G. Hopfgartner, Inorg. Chem., 2000, 39, 5059.
13 S. Rigault, C. Piguet, G. Bernardinelli and G. Hopfgartner, Angew.
Chem., Int. Ed., 1998, 27, 169.
14 (a) I. Grenthe, J. Am. Chem. Soc., 1961, 83, 360; (b) P. A. Brayshaw,
J.-C. G. Bünzli, P. Froidevaux, J. M. Harrowfield, Y. Kim and
A. N. Sobolev, Inorg. Chem., 1995, 34, 2068; (c) J. M. Harrowfield,
Y. Kim, B. W. Skelton and A. H. White, Aust. J. Chem., 1995, 48,
807; (d ) C. N. Reilley, B. W. Good and J. F. Desreux, Anal. Chem.,
1975, 47, 2110; (e) E. Huskowska and J.-P. Riehl, Inorg. Chem., 1995,
34, 5615.
1
ion. Upon excitation in the ligand ππ* state the integrated
and corrected relative intensities of the 5D0 7FJ transitions are
comparable to those measured for the solution in methanol:
0.04, 1.00, 6.98, 0.08 and 0.95 for J = 0, 1, 2, 3 and 4, respec-
tively. The very weak 5D0 7F0 transition occurs at 17 218 cmϪ1
;
it is very broad (full width at half height fwhh = 46 cmϪ1
)
and unsymmetrical on its high energy side, pointing to some-
what different environments for the EuIII ions most probably
resulting from crystal defects. The emission spectrum may again
be interpreted in terms of a severely distorted D3 symmetry
7
(cf. Table S4, ESI): the F1 sublevel is split into one singlet
and one closely spaced doublet (31 cmϪ1), while the 7F2 level is
split into one singlet and two doublets. Using the correlation
5
proposed between the energy ν of the D0 7F0 transition at
295 K and the ability of the co-ordinating atoms to produce a
nephelauxetic effect δ, ν Ϫ ν0 = CCNΣiniδi, with CCN = 1, ν0 = 17
Ϫ1
374 cm
δ
= Ϫ17.2, and δNpy = Ϫ12.1,45 we obtain ν = 17 235
CO
cmϪ1 a value which roughly matches the experimental value of
17 230 cmϪ1 for Cs3[Eu(L4 Ϫ 2H)3] (recalculated at 295 K with a
dependence of 1 cmϪ1 per 24 K).
Since L4 possesses a low-energy triplet state, it appears that it
would be suited for sensitisation of the YbIII ion,46 the energy
difference between the ligand triplet state and the metal
2F5/2 excited state being around 7000 cmϪ1. The luminescence
spectrum of the Cs3[Yb(L4 Ϫ 2H)3] compound, recorded in the
solid state at 295 K under excitation through the ligand levels,
indeed consists of an intense band centred at 10 233 cmϪ1
2
2
(fwhh = 119 cmϪ1), assigned to the F5/2 F7/2 transition.
Therefore, in spite of the small ∆E(2F5/2 Ϫ F7/2) gap, which
2
favours non-radiative deactivation processes, ligand L4 acts as a
good antenna for YbIII.
15 (a) F. Renaud, C. Piguet, G. Bernardinelli, J.-C. G. Bünzli and
G. Hopfgartner, Chem. Eur. J., 1997, 3, 1646; (b) F. Renaud,
C. Piguet, G. Bernardinelli, J.-C. G. Bünzli and G. Hopfgartner,
Chem. Eur. J., 1997, 3, 1660.
16 G. Muller, B. Schmidt, J. Jiricek, G. Hopfgartner, J. P. Riehl,
J.-C. G. Bünzli and C. Piguet, J. Chem. Soc., Dalton Trans., 2001,
2655.
17 J. B. Lamture, Z. H. Zhou, A. S. Kumar and T. G. Wensel, Inorg.
Chem., 1995, 34, 864.
18 E. R. Malinowski and D. G. Howery, Factor Analysis in Chemistry,
J. Wiley, New York, Chichester, Brisbane and Toronto, 1980.
19 H. Gampp, M. Maeder, C. J. Meyer and A. D. Zuberbühler, Talanta,
1986, 33, 943.
Conclusion
The introduction of an alkyne substituent in the 4-position
of the pyridine ring in ligand L4 is obtained with good over-
all yield. Under stoichiometric conditions, triple-stranded
helicates [Ln(L4 Ϫ 2H)]3Ϫ are formed in methanol and water,
the stability of which in the latter solvent is fairly large,
although lower than that of the [Ln(L1 Ϫ 2H)3]3Ϫ complexes,
probably because of the large electron delocalisation onto the π
system of the ligand and/or partial protonation of the terminal
amino groups. The solution structure determined in water by
paramagnetic NMR measurements shows the LnIII ions being
nine-co-ordinated with a chemical environment close to that
observed for the [Ln(L1 Ϫ 2H)3]3Ϫ complexes in the solid state.
20 P. K. Glasoe and F. A. Long, J. Phys. Chem., 1960, 64, 188.
21 C. Piguet, A. F. Williams, G. Bernardinelli, E. Moret and
J.-C. G. Bünzli, Helv. Chim. Acta, 1992, 75, 1697.
3090
J. Chem. Soc., Dalton Trans., 2001, 3084–3091

View Article Online
22 S. Petoud, T. Glanzmann, J.-C. G. Bünzli, C. Piguet, X. Qin and
R. P. Thummel, J. Lumin., 1999, 82, 69.
23 S. R. Meech and D. Phillips, J. Photochem., 1983, 23, 193.
24 H. Takalo, J. Kankare and E. Hänninen, Acta Chem. Scand., Ser. B,
1988, 42, 448.
25 C. A. Grob and H. R. Pfaendler, Helv. Chim. Acta, 1971, 54,
2060.
26 J. F. Desreux, in Lanthanide Probes in Life, Chemical and Earth
Sciences, ed. J.-C. G. Bünzli and G. R. Choppin, Elsevier Science
Publishers B. V., Amsterdam, 1989, ch. 2.
35 (a) R. M. Golding and M. P. Halton, Aust. J. Chem., 1972, 25, 2577;
(b) A. A. Pinkerton, M. Rossier and S. Stavros, J. Magn. Reson.,
1985, 64, 420.
36 B. Bleaney, J. Magn. Reson., 1972, 8, 91.
37 J. Ren and D. Sherry, J. Magn. Reson., Ser. B, 1996, 111, 178;
J. Reuben, J. Magn. Reson., 1982, 50, 233.
38 C. Platas, F. Avecilla, A. de Blas, C. F. G. C. Geraldes, T. Rodriguez-
Blas, H. Adams and J. Mahia, Inorg. Chem., 1999, 38, 3190.
39 S. Rigault, C. Piguet and J.-C. G. Bünzli, J. Chem. Soc., Dalton
Trans., 2000, 2045.
27 W. C. Wolsey, J. Chem. Educ., 1973, 50, A335.
28 (a) K. Sonogashira, Y. Tohda and N. Hagihara, Tetrahedron Lett.,
1975, 50, 4467; (b) H. A. Diek and F. R. Heck, J. Organomet. Chem.,
1975, 93, 259.
29 M. R. Lewis, D. W. Deerfield, R. A. Hoke, K. A. Koehler,
L. G. Pedersen and R. G. Hiskey, J. Biol. Chem., 1988, 263, 1358.
30 R. M. Tichane and W. E. Bennett, J. Am. Chem. Soc., 1957, 79,
1293.
40 (a) D. P. Glover-Fischer, D. H. Metcalf, T. A. Hopkins, V. J. Pugh,
S. J. Chisdes, J. Kankare and F. S. Richardson, Inorg. Chem., 1998,
37, 3026; (b) S. D. Belair, C. L. Maupin, M. W. Logue and J. P. Riehl,
J. Lumin., 2000, 86, 61.
41 Cerius-2, version 3.0, Molecular Simulations Inc., San Diego, USA,
1997.
42 J. H. Forsberg, R. M. Delaney, Q. Zhao, G. Harakas and
R. Chandran, Inorg. Chem., 1995, 34, 3705.
31 B. Perlemutter-Hayman, Acc. Chem. Res., 1986, 19, 90.
32 J. A. Peters, J. Huskens and D. J. Raber, Prog. NMR Spectrosc.,
1996, 28, 283; J. H. Forsberg, in Handbook on the Physics and
Chemistry of Rare Earths, ed. K. A. Gschneidner, Jr. and L. Eyring,
Elsevier Science Publishers B.V., Amsterdam, 1996, vol. 23, ch. 153,
p. 1ff.
33 A. D. Sherry C. F. G. C. Geraldes, in Lanthanide Probes in Life,
Chemical and Earth Sciences, ed. J.-C. G. Bünzli and G. R. Choppin,
Elsevier Science Publishers B. V., Amsterdam, 1989, ch. 4.
34 C. N. Reilley, B. W. Good and J. F. Desreux, Anal. Chem., 1975, 47,
2110.
43 Y. An, M. T. Berry and F. C. J. M. van Veggel, J. Phys. Chem. A,
2000, 104, 11243.
44 F. J. Steemers, W. Verboom, D. N. Reinhoudt, E. B. Vandertol and
J. W. Verhoeven, J. Am. Chem. Soc., 1995, 117, 9408.
45 S. T. Frey and W. DeW. Horrocks, Jr., Inorg. Chim. Acta, 1995, 229,
383.
46 C. Platas, F. Avecilla, A. de Blas, T. Rodriguez-Blas, C. F. G. C.
Geraldes, E. Tóth, A. E. Merbach and J.-C. G. Bünzli, J. Chem. Soc.,
Dalton Trans., 2000, 611.
47 I. Hemmilä, T. Ståhlberg and P. Mottram, Bioanalytical Applications
of Labelling Technologies, Wallac Oy, Turku, 2nd edn., 1995.
J. Chem. Soc., Dalton Trans., 2001, 3084–3091
3091