A simple chemical tuning of the effective concentration: Selection of single-, double-, and triple-stranded binuclear lanthanide helicates

Article abstract of DOI:10.1002/chem.200902026

The replacement of terminal 2-benzimidazol-6-carboxypyridine (two internal rotational degrees of freedom) with 2-benzimidazol-8-hydroxyquinoline (one internal rotational degree of freedom) into segmental bis-tridentate ligands in going from L2 and [L32H]²- to [L12D-2H]²- does not significantly affect the structures of the resulting binuclear lanthanide triplestranded helical complexes [Ln₂(L2)₃]⁶⁺, [Ln₂(L3-2H)₃], and [Ln₂(L12b-2H)₃] (palindromic helices, intermetallic contact distance 9 A, helical pitch 1.4 nm per turn). However, their thermodynamic assemblies are completely different in solution, as evidenced by the spectacular decrease of the effective concentrations by two orders of magnitude for [L12b-2H] ^2-. This key parameter in the [Ln₂(L12b-2H)_n] (n = 2, 3) complexes is further abruptly modulated along the lanthanide series (Ln = La to Lu), which provides an unprecedented tool for 1) tuning the number of ligand strands in the final helicates, 2) selectively coordinating lanthanides in the various complexes, and 3) controlling the ratio of lanthanide-containing polymers over discrete assemblies.

Full text of DOI:10.1002/chem.200902026

FULL PAPER
DOI: 10.1002/chem.200902026
A Simple Chemical Tuning of the Effective Concentration: Selection of
Single-, Double-, and Triple-Stranded Binuclear Lanthanide Helicates
Emmanuel Terazzi,* Laure Guꢀnꢀe, Bernard Bocquet, Jean-FranÅois Lemonnier,
Natalia Dalla Favera, and Claude Piguet*[a]
Dedicated to Professor Jean-Claude G. Bꢀnzli on the occasion of his 65th birthday
Abstract: The replacement of terminal
2-benzimidazol-6-carboxypyridine (two
internal rotational degrees of freedom)
with 2-benzimidazol-8-hydroxyquino-
line (one internal rotational degree of
freedom) into segmental bis-tridentate
ligands in going from L2 and [L3-
2H]2ꢀ to [L12b-2H]2ꢀ does not signifi-
cantly affect the structures of the re-
sulting binuclear lanthanide triple-
[Ln2(L12b-2H)3] (palindromic helices,
intermetallic contact distanceꢁ9 ꢀ,
helical pitchꢁ1.4 nm per turn). How-
ever, their thermodynamic assemblies
are completely different in solution, as
evidenced by the spectacular decrease
of the effective concentrations by two
orders of magnitude for [L12b-2H]2ꢀ.
This key parameter in the [Ln2(L12b-
2H)n] (n=2, 3) complexes is further
abruptly modulated along the lantha-
nide series (Ln=La to Lu), which pro-
vides an unprecedented tool for
1) tuning the number of ligand strands
in the final helicates, 2) selectively co-
ordinating lanthanides in the various
complexes, and 3) controlling the ratio
of lanthanide-containing polymers over
discrete assemblies.
Keywords: effective concentration ·
helical structures
· lanthanides ·
stranded
helical complexes
[Ln2ACHNUTTGREN(GNUN L3-2H)3], and
self-assembly · thermodynamics
[Ln2(L2)3]6+
,
Introduction
lated segmental ligands L2–L5—in which 1) the nature of
the donor atoms of the tridentate binding sites were some-
how varied, but 2) the two degrees of freedom produced by
internal rotations about the central pyridine ring were re-
The assembly of the first binuclear lanthanide triple-strand-
ed helicates [Ln2(L1)3]6+ (LnIII is a trivalent lanthanide)
were reported more than fifteen years ago. It relied on the
well-established concept of a judicious matching between
the binding possibilities of the ligands (segmental bis-triden-
tate and helical twist) and the stereochemical preferences of
the metals (nine-coordinate, no ligand field).[1] The subse-
quent systematic exploitation of this approach with the re-
tained—produced binuclear [Ln2(Lk)3]6+ (k=1, 2),[1,2] [Ln2-
[4]
(L3-2H)3],[3] trinuclear [Ln3(L4)3]9+
,
and tetranuclear
[Ln4(L5)3]12+[5] triple-stranded helicates (Scheme 1). The
metal-centered luminescence of these helicates and that of
the closely related system [Ln2ACTHNUTRGNEUNG
(L6-2H)3][6] gave some valua-
ble insights into the intimate mechanisms of intramolecular
f–f energy-transfer processes.[1–6] The recent decoration of
ligand L3 with water-solubilizing groups opened remarkable
[a] Dr. E. Terazzi, Dr. L. Guꢁnꢁe, B. Bocquet, Dr. J.-F. Lemonnier,
N. Dalla Favera, Prof. Dr. C. Piguet
perspectives for the use of some derivatives of [Ln2ACHTUNGTRENNUNG(L3-
2H)3] as bioprobes.[7] Because of the transparency of biolog-
ical tissue toward near-infrared (NIR) radiation, the field of
lanthanide bioprobes evolved toward the design of novel li-
gands, which were able 1) to accommodate an efficient NIR
emitter (Ln=Pr, Nd, Er, Yb), and 2) to extend the potential
excitation light sources into the visible part of the electro-
magnetic spectrum.[8] In this context, considerable efforts
have been focused on the incorporation of the fused hetero-
cyclic bidentate 8-hydroxyquinoline unit into tridentate
binding ligands L7–L10, in which an additional donor group
is connected at the 2-position of the pyridine ring
(Scheme 2).[9–12] Upon deprotonation of the phenol ring and
Department of Inorganic, Analytical and Applied Chemistry
University of Geneva, 30 quai E. Ansermet
1211 Geneva 4 (Switzerland)
Fax : (+41)22-379-6830
E-mail: Emmanuel.Terazzi@unige.ch
Claude.Piguet@unige.ch
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200902026. This includes the
synthesis of ligand L10b and statistical factors for complexes used in
multilinear least-squares fits; tables of crystallographic data and geo-
metrical analyses; tables of elemental analyses; ESIMS titrations and
thermodynamic data; schemes and figures showing the synthetic,
spectroscopic, crystallographic, and thermodynamic data; and cif files
for the crystal structures of 7 and 10.
Chem. Eur. J. 2009, 15, 12719 – 12732
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
12719

complexation with trivalent lan-
thanides, the saturated neutral
[LnACHTUNGTRENNUNG(Lk-H)3] complexes (k=7–
10) are readily formed as mix-
tures of head-to-head-to-head
(HHH, C3-symmetrical) and
head-to-head-to-tail (HHT, C1-
symmetrical) isomers. A ligand-
centered charge-transfer band
develops in the visible domain,
which can be exploited for sensi-
tizing NIR lanthanide emit-
ters.[10–12]
Though the isolation of
a
single isomer may result from
lucky crystallization processes, or
through selective interactions of
the
coordinated
phenolate
anions with cocrystallized alka-
line cations,[10–12] more reliable
strategies are to 1) connect three
parallel tridentate binding units
to
a
covalent tripod,[9,10] or
2) link two opposite tridentate
binding units to a symmetrical
spacer as found in ligands L11
and L12. The latter approach
leads to the formation of two
identical HHH nine-coordinated
building blocks related by two-
fold symmetry axes in the final
binuclear helicates [Ln
2ACHTUNGTNER(NUNG L11-
2H)3] (Ln=Eu, Yb)[13] and
[Ln2(L12a-2H)3]
(Ln=La,
Nd).[12]
The existence of neutral D3-
symmetrical triple helices for
[Ln2ACHTUNGTRENNUNG(L11-2H)3] has been estab-
lished both in the solid state
(Yb···Yb=7.77 ꢀ) and in solu-
tion. The stability constants in
anhydrous acetonitrile [equilibri-
Scheme 1. Chemical structures with numbering scheme for the central pyridine rings of the segmental multi-
tridentate ligands L1–L6. The curved arrows highlight the free rotations about the single bonds connected
to the central pyridine rings.
um
shown
in
Eq. (1);
and
log(b2E,u3,L11-2 H)=26.1(3)
log(bY2,3b,L11-2 H)=25.7(3)],[13]
al-
though similar to those found for
the charged helicates
According to the concept that any metal–ligand assembly
[Ln2(Lk)3]6+ (k=1, 2),[1,2] are alarmingly low and can be
compared with log(bE2,u3,L3-2 H)ꢁ51 (9.8<pH<10.4)[3] estimat-
ed for [L3-2H]2ꢀ in water, a solvent as competitive as aceto-
nitrile for complexation with trivalent lanthanides.[3,7]
[equilibrium shown in Eq. (2)] can be modeled with the help
of five microscopic thermodynamic parameters in Equa-
tion (3), we suspect that a thorough understanding of the
thermodynamic process leading to [Ln2(L2)3]6+, [Ln2(L3-
2H)3], and [Ln2(L12-2H)3] may deliver some clues for pro-
gramming selectivity and stability along the lanthanide
M;L
inter
series (w is the statistical factor of the assembly,[14]
f
is
M,L
m,n
3þ
3 ½L11-2 Hꢂ2ꢀ þ 2 ½LnðCH3CNÞ9ꢂ нLn2ðL11-2 HÞ3ꢂ
the absolute intermolecular affinity (including desolvation)
of one site of the ligand for the entering metal, ceff is the ef-
fective concentration used to estimate the preorganization
þ18 CH3CN
ð1Þ
Ln,L11-2 H
2,3
b
12720
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12719 – 12732

Binuclear Lanthanide Helicates
FULL PAPER
this context, the reduction of the number of rotational de-
grees of freedom per tridentate binding unit in going from
L1–L6 (n=2, see curved arrows in Scheme 1) to L7–L12
(n=1, see curved arrows in Scheme 2) is appealing, and
ligand L12 is ideally suited for elucidating the effect pro-
duced by the fusion of the terminal phenol ring with the
central pyridine ring in the 8-hydroxyquinoline moieties be-
cause the diphenylmethane spacers are identical in L1–L5
and L12.[17–19] Moreover, a preliminary NMR spectroscopy
report by Bꢃnzli and co-workers on the formation of
[La2(L12a-2H)3] confirmed the formation of the binuclear
triple helix in solution, but solubility problems and compli-
cated behaviors along the lanthanide series prevented fur-
ther characterizations except for the recording of photo-
physical data in [Nd2(L12a-2H)3].[12] These two points are
encouraging for the investigation of unprecedented thermo-
dynamic changes along the series. In this contribution, we
thus focus on the exploration of the formation of the binu-
clear triple-stranded helicates [Ln2(L12b-2H)3], with the
two following specifications in mind: 1) an analysis of the
molecular structure of [Ln2(L12b-2H)3], followed by its
comparison with [Ln2(L2)3]6+ and [Ln2(L3-2H)3], and 2) a
thermodynamic analysis of the formation of [Ln2(L12b-
2H)3] and [Ln2(L2)3]6+ along the lanthanide series.
Results and Discussion
Scheme 2. Chemical structures of the ligands L7–L12 incorporating 8-hy-
droxyquinoline units with the numbering scheme. The curved arrows
highlight the free rotations about the single bonds connecting the benz-
imidazole and the quinoline rings.
Synthesis of the ligands: The segmental ligand L12b is ob-
tained in a good global yield (48%) by means of a conver-
gent six-step synthetic procedure (Scheme 3). An improved
preparation of the precursor 3[20] starts with the O-benzyl
protection of 8-hydroxyquinaldine to give 1. A subsequent
two-step oxidation using SeO2 followed by treatment with
H2NSO3H/NaClO2 provides, successively, the aldehyde 2
and the carboxylic acid 3. The benzyl-protected ligand 7 is
then obtained by condensation between precursors 3 and
5,[21] followed by a double-reductive modified Philips cou-
pling reaction.[22] Finally, 7 is deprotected by reaction with
of the receptor for intramolecular macrocyclization reac-
M,L
MM/RT
LL/RT
j
tions (fMint,rLa =cefffinter), and uMi ,M =eꢀDE
and uLj ,L =eꢀDE
i
are the Boltzmann factors measuring intermetallic and inter-
ligand interactions in the final helicate, respectively).[15]
mzþ
M,L
m,n
m Mzþ þ n L Ð ½MmLnꢂ
DG0M,L ¼ ꢀRTlnðb
Þ
ð2Þ
ð3Þ
m,n
1
BBr3 and yields the bis-tridentate ligand L12b. Its H NMR
spectrum confirms the existence of a dynamically averaged
C2v symmetry on the NMR spectroscopic timescale (eight
signals for the aromatic protons, two pairs of enantiotopic
methylene protons in 1:2 ratio, and two signals for methyl
groups) as previously reported for the analogous ligand
L12a.[12] The parent monotopic tridentate ligand L10b
(Scheme 2) is prepared by using the same synthetic strategy
in three steps from 3 and 4[21] (yield 42%; see Scheme S1 in
the Supporting Information).
For the two uncomplexed ligands L10b and L12b, the
lack of a nuclear Overhauser enhancement (NOE) effect
between the ethyl residues of the benzimidazole ring and
the protons connected at the 3-position of the quinoline ring
(see Scheme 2 for numbering) agrees with a trans arrange-
ment of the N-donor atoms borne by the adjacent aromatic
rings (Scheme 3). Slow evaporation of a solution of the pre-
cursor ligand 7 in CH2Cl2/MeOH gives colorless prisms, for
Y
Y
L;L=RT
M,L
m,n
M,L M,L mn
m,n
mnꢀnꢀmþ1
M;M=RT
i
j
b
¼w ðfinterÞ
ðceffÞ
eꢀDE
eꢀDE
i
j
Y
Y
¼ w ðfinter
Þ
ðceffÞ
uMi ;M
uLj ;L
M,L M,L mn
mnꢀnꢀmþ1
m,n
i
j
Among the different candidates for rationalizing the drastic
decrease in stability by twenty-five orders of magnitude in
going from [Ln2(L3-2H)3] (water) to [Ln2(L11-2H)3] (aceto-
nitrile), the lower polarity of acetonitrile is a crucial point,
as it reduces the entropy gain produced by the charge neu-
tralization accompanying the complexation processes.[16]
However, some variations of the effective concentration, ceff,
which measures the degree of preorganization of the recep-
tor for the specific intramolecular macrocyclization process-
es operating in the formation of triple-stranded helicates,
may also significantly contribute to their final stabilities. In
Chem. Eur. J. 2009, 15, 12719 – 12732
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12721

E. Terazzi, C. Piguet et al.
droxyquinoline ring, the O-
benzyl-protected
quinoline
atom is forced to adopt a con-
strained cis planar conforma-
tion with the adjacent N-quin-
oline atom, which contrasts
with the more flexible trans–
trans arrangement found for
tridentate 2,6-disubstituted pyr-
idine units in ligands L1–
L6.[2–5] Bond lengths and bond
angles are standard (Tables S1–
S3 in the Supporting Informa-
tion),[23] and the ligand possess-
es a pseudo-twofold axis along
the
c direction and passing
through C26. Two intermolecu-
lar stacking interactions can be
detected in the unit cell (Fig-
ure S1 and Table S3 in the Sup-
porting Information).
Finally,
the
quantitative
analysis of the helicity of the
diphenylmethane spacer, de-
fined as a five-atom crooked
line (C19-C20-C26-C46-C45 in
7), by using Equation (4)[24]
gives a helicity index (H) of
0.13, which is diagnostic for a
negligible helical preorganiza-
tion of the ligand strand (L=
3.51 ꢀ is the end-to-end dis-
Scheme 3. Reagents: i) benzyl bromide, K2CO3, KI (cat.), acetone; ii) SeO2, dioxane; iii) H2NSO3H, NaClO2,
H2O/THF; iv) EtNH2/DMSO; v) (CH2O)n/HCl (37%); vi) compound 3, SOCl2, DMF (cat.), CH2Cl2, then 5,
CH2Cl2, diisopropylethylamine; vii) Fe, H2O/EtOH, HCl (37%); viii) BBr3 (1m) in CH2Cl2.
which the X-ray crystal structure, shown in Figure 1, con-
firms that the tridentate binding units adopt the expected
trans conformation, which minimizes the global dipole mo-
mentum (i.e., the coordinating N-benzimidazole atom is
trans to the N-quinoline atom with respect to the interannu-
tance of the helix taken as the C19···C45 contact distance;
A=0.215 ꢀ2 is the area of the subtended figure produced by
the projection of the five atoms onto a plane perpendicular
to the heliACTHNUTRGNEUGcN al axis defined by the line passing through two
terminal atoms of the chain; and D=5.82 ꢀ is the total
lar C C bond). Because of the fused character of the hy-
length of the crooked line).[25]
ꢀ
pffiffiffiffiffi
LA
D3
ð4Þ
H ¼ 6 3p
Synthesis and structure of the
complexes [Ln2(L12b-2H)3] in
the solid state: Reaction of the
C2v-symmetrical bis-tridentate
ligand L12b (3 equiv) with
Ln(OTf)3·xH2O (2 equiv; Ln=
ꢀ
La, Nd, Eu; OTfꢀ =CF3SO3
=
triflate; x=1–3) in acetonitrile
containing K2CO3, followed by
precipitation
ether, yields pure red-orange
binuclear complexes
with
diethyl
[Ln2(L12b-2H)3]·xH2O
(Table S4 in the Supporting In-
formation). More than two
Figure 1. View of the crystal structure of the precursor ligand 7 with numbering scheme. Ellipsoids are repre-
sented at the 30% probability level.
12722
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12719 – 12732

Binuclear Lanthanide Helicates
FULL PAPER
hundred crystallization attempts provided some extremely
fragile red crystals, from which an X-ray crystal structure of
limited quality (see the Experimental Section) could be
eventually obtained (Figure 2, Table 1; and Tables S5–S7 in
the Supporting Information).
Two independent neutral binuclear triple-stranded helical
complexes possessing opposite helicities A-[Nd2(L12b-2H)3]
(multiplicity: 8) and B-[Nd2(L12b-2H)3] (multiplicity: 4) are
found in the unit cell, together with interstitial solvent mole-
cules, thus leading to the complete chemical formula {A-
[Nd2(L12b-2H)3]}2·{B-[Nd2(L12b-2H)3]}·8CH3CN·4H2O
(10) with Z=4. Although B-[Nd2(L12b-2H)3] is located on
a crystallographic twofold axis passing through C19e, the
metric of the two independent complexes is very similar
(Table 1; Tables S5–S7 in the Supporting Information); they
show pseudo-D3 symmetrical binuclear triple-stranded heli-
cates. The two metallic cations NdIII are nine-coordinated in
pseudo-tricapped trigonal prismatic sites, with each HHH
site being defined by three oxygen atoms of the terminal
ꢀ
phenol rings (upper tripod, average Nd Ophenol =2.40(1) ꢀ),
three capping nitrogen atoms of the quinolines (average
ꢀ
Nd Nquin =2.65(1) ꢀ), and three nitrogen atoms of the ben-
ꢀ
zimidazole rings (lower tripod, average Nd Nbzim
=
2.72(1) ꢀ; Table 1, Figure 2; and Scheme S2 in the Support-
ꢀ
ꢀ
ing Information). The Nd O and the Nd N bond lengths
are comparable to those reported for the mononuclear com-
plex HHT-[Nd(L10-H)3] (R=Me in Scheme 2; average
ꢀ
ꢀ
Nd Ophenol =2.406(8) ꢀ, average Nd Nquin =2.64(2) ꢀ, and
average Nd Nbzim =2.73(6) ꢀ).[12] This logically leads to the
ꢀ
calculation of similar nine-coordinate ionic radii randomly
distributed around the values of 1.163 ꢀ that are expected
for unconstrained nine-coordinate NdIII.[26] We thus deduce
that 1) the different relative orientations of the three 2-
benzim
NdIII (HHH or HHT), and 2) the incorporation of the
benzimidazole ring into the diphenylmethane spacer in
ACHTUNGTRENiNGNU dazol-8-hydroxyquinoline binding units bound to
AHCTUNGTRENNUNG
[Nd2(L12b-2H)3] has no major structural effect in the final
complexes. However, the average bond valences,[27,28] which
reflects the strength of the Ln–donor interactions,[29] provide
evidence for stronger interactions for the terminal oxygen
atoms in [Nd2(L12b-2H)3] at the cost of some weaker inter-
actions with the nitrogen atoms of the benzimidazole rings,
thus maintaining the bond-valence sum within the accepta-
ble range of VNd =3.0ꢃ0.2 (Table 1).[30] Comparison with
[Tb2(L3-2H)3], in which the terminal-fused hydroxyquino-
line rings are replaced with pyridine–carboxylate units
(Schemes 1 and 2), shows the opposite trend with a smaller
difference between nLn,O and nLn,Nbzim (Table 1). An analysis
Figure 2. Perspective view with numbering scheme of the molecular
structure of A-[Nd2(L12b-2H)3] in the crystal structure of 10. Ellipsoids
are represented at 30% probability level. The three strands are repre-
sented with different grayscale intensities (the respective strands are
named a, b, and
c for A-[Nd2(L12b-2H)3], and d, e, and f for
B-[Nd2(L12b-2H)3]).
Table 1. Ln Ln distances, average Ln N and Ln O bond lengths [ꢀ], bond valences (nLn,j),[a] bond-valence sums (VLn),[b] and geometric analysis of the
coordination polyhedra in A-[Nd2(L12b-2H)3] and B-[Nd2(L12b-2H)3] helicates and related complexes.
ꢀ
ꢀ
ꢀ
[c]
[c]
[c]
[c]
ꢀ
ꢀ
Ln O
Compound
Ln N
Ln···Ln
nLn,N
nLn,O
VLn
Ref.
Bzim[d]
Quin[d]
Bzim[d]
Quin[d]
A-[Nd2(L12b-2H)3]
Nd2
Nd3
Nd1
Tb1A
Tb2A
Tb1B
Tb2B
Tb1
2.72(6)
2.71(4)
2.73(3)
2.60(1)
2.62(1)
2.61(1)
2.60(4)
2.56(1)
2.59(5)
2.66(3)
2.63(1)
2.65(1)
2.55(1)
2.56(2)
2.54(2)
2.52(1)
2.57(3)
2.59(3)
2.41(1)
2.41(2)
2.39(1)
2.37(2)
2.36(1)
2.37(5)
2.39(3)
2.36(4)
2.36(3)
9.021(2)
–
9.047(2)
8.829(1)
0.25(3)
0.25(3)
0.24(2)
0.285(5)
0.31(2)
0.271(9)
0.353(4)
0.31(1)
0.29(4)
0.29(2)
0.314(8)
0.30(4)
0.32(1)
0.266(9)
0.33(1)
0.28(4)
0.31(3)
0.29(2)
0.413(6)
0.42(3)
0.44(5)
0.39(3)
0.394(7)
0.38(5)
0.37(3)
0.39(4)
0.39(3)
2.84
2.96
2.94
2.97
2.92
2.95
2.99
3.03
2.91
this work
this work
this work
[3a]
B-[Nd2(L12b-2H)3]
A-[Tb2(L3-2H)3]
–
B-[Tb2(L3-2H)3]
[Tb2(L2)3]6+
Ln,jꢀdLn,j
9.069(1)
–
9.06(3)
–
[3a]
[2a]
Tb2
[a] nLn,j =e[(R
)/b], in which dLn,j is the Ln–donor atom j distance. The bond-valence parameters RLn,N and RLn,O are taken from ref. [28] and b=
P
0.37 ꢀ.[27] [b] VLn
=
nLn,j
.
[c] Each value is the average of three bond lengths and the number in parentheses corresponds to the standard deviation
[28]
j
of the average. The original uncertainties affecting each bond length are given in Table S5 in the Supporting Information. [d] Bzim=benzimidazole and
Quin=quinoline.
Chem. Eur. J. 2009, 15, 12719 – 12732
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12723

E. Terazzi, C. Piguet et al.
of the geometries, shapes and helicities of these two com-
plexes shows very similar characteristics (Tables S8 and S9,
and Figures S2 and S3 in the Supporting Information). Con-
sequently, the Nd···Nd contact distance in the [Nd2(L12b-
2H)3] (9.02–9.05 ꢀ) is almost identical to those found in
[Tb2(L3-2H)3] and [Tb2(L2)3]6+ (Table 1), thus leading to
approximate palindromic helices (lengthꢁ1.4 nm, diameter
ꢁ1.1 nm, pitchꢁ1.4 nm per turn).
charge-transfer (ILCT) transitions (400–500 nm),[12] which
result from the complexation of the tridentate N2Oꢀ binding
units to Ln3+ (Figure 3). We observe two end points around
Ln/L12b=0.67 and Ln/L12b=1.0, followed by a smooth
evolution for Ln/L12b>1.0, which is almost negligible for
large lanthanides (Ln=La, Nd, Eu; Figure 3a) but very sig-
nificant for small lanthanides (Ln=Y, Lu; Figure 3b).
Factor analyses[32] suggest the formation of, respectively,
three (large LnIII) or four (small LnIII) absorbing species
during the titration processes, and the spectrophotometric
data can be satisfyingly fitted with the equilibria shown in
Equations (5) and (6) for Ln=La, Nd, Eu, and Equa-
tions (5)–(7) for Ln=Y, Lu, by using nonlinear least-squares
techniques.[33]
Interestingly, reaction of the L12b (3 equiv) with
Ln(OTf)3·xH2O (2 equiv; Ln=Y, Lu ; x=1–3) in acetonitrile
containing K2CO3 failed to give pure binuclear complexes
[Ln2(L12b-2H)3]·xH2O, which are systematically contami-
nated with 15–25% of the highly insoluble complexes
[Ln2(L12b-2H)2(OTf)2] that possess Ln/L12=2:2 stoichiom-
etry (Table S4 in the Supporting Information).
Ln;L12b
2;3
3 ½L12 b-2 Hꢂ2ꢀ þ 2 Ln3þ Ð ½Ln2ðL12 b-2 HÞ3ꢂ
2 ½L12 b-2 Hꢂ2ꢀ þ 2 Ln3þ Ð ½Ln2ðL12 b-2 HÞ2ꢂ
b
b
ð5Þ
ð6Þ
Speciation, stability, and structure of the complexes
[Ln2(L12b-2H)3] in solution: The faint solubility of both
free ligand L12b and of its complexes [Ln2(L12b-2H)3] in
water or in polar organic solvent represents a severe handi-
cap for the characterization of the assembly processes oper-
ating in solution. After considerable effort, we found that ti-
trations of the deprotonated ligand [L12b-2H]2ꢀ (obtained
from reaction of L12b with 2 equiv of (nBu)4NOH) with in-
creasing quantities of Ln(OTf)3·xH2O (Ln=La, Nd, Eu, Y,
Lu; x=1–3) can be performed at 10ꢀ4 m in CH2Cl2/CH3OH
(1:1), or up to 2ꢄ10ꢀ3 m in DMSO, although the pure depro-
tonated ligand rapidly (within one hour) decomposes in the
latter solvent. Monitoring the titrations with electrospray
ionization mass spectrometry (ESIMS) for Ln/L12b in the
range 0.1 to 2.0 displays similar results in both solvents and
shows the successive formation of three charged complexes
Ln;L12b
2;2
2þ
Ln;L12b
2;1
4þ
½L12 b-2 Hꢂ2ꢀ þ 2 Ln3þ Ð ½Ln2ðL12 b-2 HÞꢂ
b
ð7Þ
Ln;L12b
2;3
The associated cumulative constants log(b
)
and
log(b2L;n2;L12b) collected in Table 2 are of the same magnitude
as those previously reported for [Ln2(L11-2H)3] and
[Ln2(L11-2H)2]2+ in acetonitrile (Ln=Eu, Yb),[13] but we
detect some selectivity with a preference for the fixation of
mid-range lanthanides (Table 2; Figure S5 in the Supporting
Information).
Interestingly, the stabilities of [Ln2(L12b-2H)3] and
[Ln2(L12b-2H)2]2+ for Ln=Nd and Lu are comparable, but
the possibility for the smaller lanthanide to form the stable
[Lu2(L12b-2H)]4+ complex limits the formation of the triple
helix (Figure 4), which partly explains the difficulties en-
countered in isolating pure neutral complexes along the
second part of the lanthanide series.
[Ln2(L12b-2H)3+2H]2+
2H)2]2+
(Ln/L12b=0.67),
[Ln2(L12b-
[Ln2(L12b-
(Ln/L12b=1.0),
and
2H)(OTf)3(H2O)4(CH3OH)4]+ (Ln/L12b=2.0), together
with some adducts of these complexes with either solvent
molecules or counterions (Figure S4 and Tables S10–S14 in
the Supporting Information). For the large lanthanides
(Ln=La, Nd, Eu), the ESIMS spectra recorded for Ln/
L12bꢄ1.0 were dominated by the signal of the doubly pro-
tonated complex [Ln2(L12b-2H)3+2H]2+ (Figure S4a),
which is probably obtained by the fixation of protons onto
the terminal negatively charged phenol tripods of the triple
helix, as previously described for K+ in the crystal structure
of HHH-[Er(L8-H)3+K]+.[10b] Further addition of metal
(Ln/L12bꢅ1.0) yields the formation of [Ln2(L12b-2H)2]2+
at the expense of [Ln2(L12b-2H)3+2H]2+. For the smaller
lanthanides (Ln=Y, Lu), the competition between
[Ln2(L12b-2H)3+2H]2+ and [Ln2(L12b-2H)2]2+ is more
pronounced (Figure S4b), whereas in an excess of metal, we
detect signals diagnostic of the formation of an unsaturated
2:1 complex [Ln2(L12b-2H)(OTf)3(H2O)4(CH3OH)4]+, for
which the ESIMS response is expected to be weak.[31]
1H NMR spectroscopic data are especially hard to obtain
because of the poor sensitivity of this technique, which usu-
ally requires concentrations in the millimolar range. Solubil-
ity of both ligands and complexes in CD2Cl2/CD3OH (1:1) is
1
not sufficient, but reliable H NMR spectra can be recorded
for saturated solutions of [Ln2(L12b-2H)3] in CD3CN with
the larger lanthanides (Ln=La, Nd, Eu, Y), whereas titra-
tions cannot be envisioned because of precipitation of
1
[Ln2(L12b-2H)2](OTf)2 in an excess of metal. The H NMR
spectra of both diamagnetic (Ln=La, Y; Figure 5) and fast-
relaxing paramagnetic (Ln=Nd, Eu; Figure S6 in the Sup-
porting Information) [Ln2(L12b-2H)3] complexes display
eight signals for the aromatic protons, which are diagnostic
for D3 (helical) or C3h symmetry (side-by-side).[34] The con-
comitant observation of singlets (i.e., A2 spin systems) for
the enantiotopic methylene protons of the diphenylmethane
spacer (H9,9’ in Figure 5) requires three twofold axes per-
pendicular to the threefold axis and thus confirms the for-
Parallel spectrophotometric titrations of [L12b-2H]2ꢀ
with Ln(OTf)3·xH2O performed at 2ꢄ10ꢀ4 m in CH2Cl2/
CH3OH (1:1) show a complicated variation of both ligand-
centered p!p* transitions (250–380 nm) and intraligand
mation of the expected D3-symmetrical triple helix.[35]
A
complete assignment is obtained by means of 2D-COSY and
2D-NOESY spectroscopies for the diamagnetic complexes,
although some ambiguities remain for the paramagnetic an-
12724
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12719 – 12732

Binuclear Lanthanide Helicates
FULL PAPER
Figure 3. Variation of the absorption spectra and corresponding variations of molar extinctions at seven different wavelengths observed for the spectro-
photometric titrations of [L12b-2H]2ꢀ (2ꢄ10ꢀ4 m in CH2Cl2/MeOH 1:1) with a) Nd(OTf)3 and b) Y(OTf)3 (Ln/L12b=0.1–3.0).
alogues since the increased nuclear relaxation severely
limits the use of these two-dimensional techniques (Fig-
ure S6 in the Supporting Information).
placement of the D3-symmetrical [Ln2(L12b-2H)3] helicate
with [Ln2(L12b-2H)2]2+ existing as a mixture of double-
stranded (D2 symmetry characterized by enantiotopic H9,9’
and diastereotopic H10,10’ methylene protons)[34] and side-
by-side (C2h symmetry characterized by two diastereotopic
pairs of methylene protons for H9,9’ and H10,10’,
Scheme 4a),[34] a situation previously documented for related
binuclear lanthanide complex with bis-bidentate Schiff base
ligands.[36] The ultimate complex [Ln2(L12b-2H)]4+ displays
1H NMR spectroscopic singlet signals for all methylene
probes H9,9’ and H10,10’, which is in line with an average
extended planar C2v symmetry of the ligand strand
(Scheme 4b) previously established for the related com-
plexes [Ln2(L2)(NO3)6].[25]
In the more coordinating [D6]DMSO solvent, the stability
constants are severely reduced and we observe a significant
amount of free ligand in slow exchange with [Ln2(L12b-
2H)3] for Ln/L12b=0.67 and a total ligand concentration of
2ꢄ10ꢀ3 m (Ln=La, Nd, Eu, Y, Lu; Figures S7–S11 in the
Supporting Information). However, the addition of an
excess of metal does not produce immediate precipitation
1
and we can record the H NMR spectroscopic signatures for
the successive formation of [Ln2(L12b-2H)2]2+ and
[Ln2(L12b-2H)]4+. For large diamagnetic (Ln=La) or para-
magnetic (Ln=Nd, Eu) lanthanides, the broadened signals
due to intermediate chemical exchange on the NMR spec-
troscopic timescale, combined with lanthanide-induced re-
laxation, prevent a detailed analysis. On the other hand, the
1H NMR spectroscopic titrations with small diamagnetic cat-
ions (Ln=Y, Lu; Figures S10 and S11) show the stepwise re-
Thermodynamic modeling and rationalization of the forma-
tion of the complexes [Ln2(L12b-2H)n](6–2n)+ (n=1–3) in so-
lution: The application of the site-binding model [Eq. (3)] to
the complexation equilibria in Equations (5)–(7) yields
Chem. Eur. J. 2009, 15, 12719 – 12732
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12725

E. Terazzi, C. Piguet et al.
Table 2. Experimental and computed cumulative formation constants
Ln;L10b
1;n
Ln;L12b
2;n
logðb
Þ and logðb
Þ obtained by spectrophotometry according to
the equilibria in Equations (5)–(7) and (11)–(13) for the complexes
[Ln2(L12b-2H)n](6ꢀ2n)+ and [Ln(L10b-2H)n](3ꢀ2n)+ (Ln=La, Nd, Eu, Y,
and Lu; CH2Cl2/CH3OH 1:1, 298 K).
LnIII
La
Nd
Eu
Y
Lu
CN¼9
R
[ꢀ][a]
1.216
1.163
1.120
1.075
1.032
Ln
Ln;L12b
logðb
logðb
logðb
logðb
logðb
logðb
AFLn
Þ
exptl[b]
23.1(6) 24.8(5) 27.0(8) 26.5(5) 24.0(4)
23.1
–
17.0(5) 18.7(4) 20.2(7) 20.9(4) 18.6(3)
17.0
–
ꢄ10
12.9
–
20.3(5) 19.7(5) 20.1(6) 23.1(9) 23.5(9)
20.3
–
14.8(4) 14.2(4) 14.5(5) 16.0(9) 16.6(6)
14.9
–
8.3(3)
8.2
–
0.0023
–
2;3
calcd1[c]
calcd2[d]
exptl[b]
24.8
–
27.0
–
26.8
26.5
24.3
24.0
Ln;L12b
2;2
Þ
Þ
Þ
Þ
Þ
calcd1[c]
calcd2[d]
exptl[b]
18.7
–
ꢄ11
13.8
–
20.2
–
ꢄ12
14.8
–
20.4
20.9
18.1
18.6
Ln;L12b
2;1
13.0(3) 11.7(2)
13.2
13.0
calcd1[c]
calcd2[d]
exptl[b]
12.0
11.7
Ln;L10b
1;3
calcd1[c]
calcd2[d]
exptl[b]
19.7
–
20.1
–
23.4
23.1
23.7
23.5
Ln;L10b
1;2
calcd1[c]
calcd2[d]
exptl[b]
14.1
–
7.6(6)
7.7
14.5
–
7.9(3)
7.9
15.6
15.8
8.0(6)
7.9
8.2
16.3
16.5
8.7(7)
8.5
8.8
Ln;L10b
1;1
calcd1[c]
calcd2[d]
calcd1[c]
calcd2[d]
–
–
[e]
0.0017
–
0.0006
–
0.0171
0.0050
0.0180
0.0025
[a] Ionic radii for nine-coordinate trivalent lanthanides.[26] [b] The quoted
errors correspond to those estimated during the nonlinear least-square
fits. [c] Computed with Equations (8)–(10) and (14)–(16) and parameters
of Table 3 (first fitting process). [d] Computed with Equations (8)–(9)
and (14)–(17) and parameters of Table 3 (second fitting process).
[e] Willcott agreement factor; see text and references.[38]
Equations (8)–(10), in which the three experimentally acces-
Ln,L12b
2,1
sible thermodynamic constants bL2,n3,L12b, bL2,n2,L12b, and b
Figure 4. Computed speciation (% ligand) for the titration of 2ꢄ10ꢀ4
m
[L12b-2H]2ꢀ with a) Nd(OTf)3 and b) Lu(OTf)3 (stability constants are
require a minimal set of four microscopic parameters for
Ln,L12b
their modeling (f
is the affinity, including desolvation,
taken from Table 2, CH2Cl2/MeOH 1:1).
N2O
of LnIII for the tridentate binding site in [L12b-2H]2ꢀ; uLn,Ln
and uL,L are the average Boltzmann factors for intramolecu-
lar homo-component interactions operating within the final
assemblies; and ceff is the average effective concentration
characterizing the macrocyclization processes responsible
for the formation of discrete double-stranded or triple-
stranded helicates), together with a statistical factor calculat-
ed by using the method of symmetry numbers, first intro-
duced by Benson (detailed calculations are given in the Sup-
porting Information).[14,37]
ants b1L;n3;L10b, b1L;n2;L10b, and b1L;n1;L10b, which correspond to the
equilibria in Equations (11)–(13) after nonlinear least-
squares fits of the data (Table 2).[33]
Ln;L10b
1;3
3 ½L10 b-Hꢂꢀ þ Ln3þ Ð ½LnðL10 b-HÞ3ꢂ
2 ½L10 b-Hꢂꢀ þ Ln3þ Ð ½LnðL10 b-HÞ2ꢂþ
b
ð11Þ
ð12Þ
ð13Þ
Ln;L10b
1;2
b
Ln;L10b
b
1;1
½L10 b-Hꢂꢀ þ Ln3þ Ð ½LnðL10 b-HÞꢂ
2þ
Ln;L12b
2;3
Ln;L12b
N2O
6
2
6
b
b
b
¼ 96ðf
¼ 72ðf
¼ 36ðf
Þ ðceffÞ ðuL;LÞ ðuLn;Ln
Þ
ð8Þ
ð9Þ
To limit the number of adjustable microscopic parameters,
we have set 1) uLH,HL =uLH,TL =uL,L, which implies statistical dis-
tributions of the microspecies HH-[Ln(L10b-H)2]+/HT-
Ln;L12b
2;2
Ln;L12b
N2O
4
4
Þ ðceff ÞðuL;LÞ ðuLn;Ln
Þ
Ln;L12b
2;1
Ln;L12b
N2O
2
Þ ðuLn;Ln
Þ
ð10Þ
[Ln(L10b-H)2]+ =1.0
and
HHH-[Ln(L10b-H)3]/HHT-
ꢀ
[Ln(L10b-H)3]=0.33 (justified by the similar Nd N and
ꢀ
The collection of a sufficient number of experimental stabili-
ty constants forced us to perform additional spectrophoto-
metric titrations of the simple tridentate ligands [L10b-H]ꢀ
with Ln(OTf)3·xH2O (2ꢄ10ꢀ4 m in CH2Cl2/CH3OH (1:1),
298 K); these eventually give three additional macroconst-
Nd O bond lengths and bond valences found in HHT-
[Nd(L10b-H)3] and (HHH)-[Nd2(L12b-2H)3] in the solid
Ln,L10b
N2O
Ln,L12b
(i.e., the tridentate binding
N2O
state), and 2) f
=f
units in [L10b-H]ꢀ and [L12b-H]2ꢀ have similar affinities
for LnIII). Application of the site-binding model [Equa-
12726
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12719 – 12732

Binuclear Lanthanide Helicates
FULL PAPER
Figure 5. 1H NMR spectra with numbering scheme for diamagnetic triple-
stranded helicates a) [La2(L12b-2H)3] and b) [Y2(L12b-2H)3] (CD3CN,
298 K).
tion (3)] eventually gives Equations (14)–(16).
Ln;L10b
1;3
Ln;L12b
N2O
3
3
b
¼ 16ðf
¼ 12ðf
Þ ðuL;L
Þ
Þ
ð14Þ
Ln;L10b
1;2
Ln;L12b
N2O
2
b
Þ ðuL;L
ð15Þ
ð16Þ
Scheme 4. Schematic representations of the solution structures estab-
lished by 1H NMR spectroscopy for a) [Ln2(L12b-2H)2]2+ existing as a
mixture of helical (D2 symmetry) and side-by-side (C2h symmetry)
double-stranded complexes, and b) [Ln2(L12b-2H)]4+ (average C2v sym-
metry on the NMR spectroscopic timescale). The coordination spheres of
LnIII have not been saturated with solvent molecules for the sake of clari-
ty.
Ln;L10b
1;1
Ln;L12b
N2O
b
¼ 6ðf
Þ
Simultaneous multilinear least-squares fits of Equations (8)–
(10) and (14)–(16) for each lanthanide yield the sets of four
microscopic thermodynamic descriptors collected in Table 3
(rows 1–8) and shown in Figure 6a, which satisfyingly repro-
duce the experimental data, as ascertained by the small
computed values of the Willcott agreement factors (0.0006ꢄ
AFLn ꢄ0.0180, Table 2).[38]
However, the most striking point concerns the very low
effective concentrations[39] (ꢀ9ꢄlog(ceff)ꢄꢀ6), which is at
least two orders of magnitude smaller than 1) log(ceff)=
ꢀ4.1(4) reported for [Lu2(Lk)n]6+ in acetonitrile (k=1, 2,
Scheme 1; n=2, 3),[19] and 2) log(ceff)=ꢀ0.3(4), estimated
from thermodynamic and kinetic data collected for
[Eum(L3-2H)n](3mꢀ2n)+ in water at pH 6.15.[15] Moreover,
log(ceff) in [Ln2(L12b-2H)n](6ꢀ2n)+ abruptly decreases in
going from Ln=Eu to Ln=Y, which translates into an in-
The main driving force of the complexation processes re-
sults from intermolecular metal–ligand connection process-
es, which do not significantly vary along the lanthanide
Ln,L12b
inter
series (ꢀ44ꢄDG
ꢄꢀ39 kJmolꢀ1). The interligand in-
teractions also poorly change along the lanthanide series,
but contribute to a global destabilization (DEL,L ꢅ0) when
the metals are stepwise saturated with tridentate binding
units. The concave behavior of DELn,Ln, which shows a mini-
mum intermetallic repulsion for mid-range LnIII (Figure 6a),
is responsible for the parallel extra stabilization of the binu-
clear [Ln2(L12b-2H)n](6ꢀ2n)+ complexes (n=2, 3) in the
middle of the series (Table 2; Figure S5 in the Supporting
Information).
version of the sign of the free-energy change characterizing
Ln,L12b
inter
intramolecular macrocyclization processes (DG
<0 is
favorable for large LnIII, but DG
>0 is unfavorable for
Ln,L12b
intra
small LnIII; Figure 6a). Although this factor may explain the
limited stability of macro(bi)cyclic double-stranded and
triple-stranded binuclear helicates with respect to the noncy-
Chem. Eur. J. 2009, 15, 12719 – 12732
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12727

E. Terazzi, C. Piguet et al.
Table 3. Fitted microscopic thermodynamic parameters for [Ln(L10b-H)n](3n)+ and [Ln2(L12b-2H)n](6ꢀ2n)+
(Ln=La, Nd, Eu, Y, and Lu; CH2Cl2/CH3OH 1:1, 298 K).[a]
huge decrease of log(ceff) in
going from L2 (Ln=Lu)[19] or
Fitted parameters (first fit)[b]
La
Nd
Eu
Y
Lu
[L3-2H]2ꢀ
(Ln=Eu)[15]
to
log(fLNn,L12b
)
7.44(6)
6.87(5)
7.14(2)
7.1(4)
7.7(3)
[L12b-2H]2ꢀ (Ln=La, Nd, Eu,
Y, Lu) to the replacement of
flexible terminal pyridine–car-
boxy(amide or late) groups
with fused ten-membered rigid
2 O
Ln,L
[c]
DG
[kJmolꢀ1
]
ꢀ42.4(4)
ꢀ6.8(1)
ꢀ3.6(6)
ꢀ1.09(8)
6.2(5)
ꢀ39.2(3)
ꢀ6.4(1)
ꢀ2.7(5)
ꢀ0.70(6)
4.0(4)
ꢀ40.7(1)
ꢀ5.83(4)
ꢀ7.5(2)
ꢀ0.86(2)
4.9(1)
ꢀ41(2)
ꢀ8.5(8)
8(3)
ꢀ44(2)
ꢀ8.9(8)
7(3)
inter
log(ceff
)
Ln,L12b
intra
[d]
DG
[kJmolꢀ1
]
logðuL;L
Þ
ꢀ0.3(4)
ꢀ0.2(4)
DEL,L [kJmolꢀ1
]
2(2)
1(2)
[e]
log(uLn,Ln
DELn,Ln [kJmolꢀ1
)
ꢀ3.5(2)
ꢀ1.5(2)
ꢀ0.99(6)
ꢀ2.5(8)
15(5)
ꢀ4.9(8)
28(5)
aromatic
8-hydroxyquinoli-
[f]
]
20(1)
9(1)
5.6(4)
nates. To substantiate this hy-
pothesis, we have repeated this
complete thermodynamic anal-
ysis for the formation of
[Ln2(L2)n]6+ (n=1–2)[2b] and
[Ln(L13)n]3+ (n=1–3) in ace-
tonitrile along the lanthanide
series. Equations (8), (9), and
(14)–(16) still apply for fitting
the experimental stability con-
stants (Table S15 in the Sup-
porting Information) to give
the microscopic parameters
shown in Figure 6b (Table S16
Fitted parameters (second fit)[g]
La
Nd
Eu
Y
Lu
log(fLNn,L12b
)
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
7.4(2)
7.99(8)
ꢀ45.6(5)
ꢀ9.5(5)
8.6(8)
2 O
[c]
DGLn,L [kJmolꢀ1
]
ꢀ42(1)
ꢀ9.1(4)
10(2)
ꢀ0.1(2)
logðicnteefrfÞ
Ln,L12b
intra
[d]
DG
[kJmolꢀ1
]
logðuL;L
Þ
ꢀ0.56(9)
DEL,L [kJmolꢀ1
]
1(1)
3.2(6)
[e]
log(uLn,Ln
)
ꢀ1.1(5)
6(3)
ꢀ3.4(3)
19(2)
DELn,Ln [kJmolꢀ1
]
[f]
Ln;Ln
logðu
Þ
ꢀ3.4(4)
ꢀ5.8(2)
single strand
[kJmolꢀ1
]
19(2)
33(1)
Ln,Ln
single strand
[e]
DE
[a] The uncertainties correspond to those found during the multilinear least-square fits. [b] First fit: fLNn,L12b, ceff
,
2O
Ln,L12b
intra
uL,L
,
uLn,Ln by using Equations (8)–(10) and (14)–(16). [c] DG
=ꢀRTln(fLNn,L12b). [d] DG
=
Ln,L12b
inter
2O
ꢀRTln(fNLn,L12bceff). [e] DEL,L =ꢀRTln(uL,L). [f] DELn,Ln =ꢀRTln(uLn,Ln). [g] Second fit: fNLn,L12b, ceff, uL,L, uLn,Ln
,
2 O
2O
Ln,Ln
u
by using Equations (8)–(9) and (14)–(17).
single strand
clic [Ln2(L12b-2H)]4+ complex, particularly for small LnIII,
it is not sufficient to justify the nondetection of the latter
complexes for large LnIII (compare experimental and com-
Ln;L12b
2;1
puted (italic) values of b
for Ln=La, Nd, Eu in
Table 2). Although the quantitative dissection of DELn,Ln
into the opposite contributions of intramolecular coulombic
repulsion and favorable solvation energies is beyond the
scope of this already massive contribution,[19] we expect that
DELn,Ln becomes more repulsive at longer distances (a very
counterintuitive result from ref. [19]), and we have also
demonstrated that the formation of the single-stranded com-
plex [Ln2(L2)(NO3)3] is accompanied by the planarization of
the ligand and an increase of approximately 30% of the in-
in the Supporting Information). Despite the use of two dif-
ferent solvents for imperious solubility reasons, we have
indeed selected two systems (CH2Cl2/CH3OH (1:1) and
CH3CN) with similar polarities, which allow a direct com-
parison of the different parameters. We immediately notice
that the absolute magnitude of the various thermodynamic
parameters are similar, whereas their variation along the
lanthanide series differ in both systems (Figure 6). However,
the larger and most striking discrepancy unambiguously con-
cerns the gradual increase of the effective concentration
tramolecular intermetallic distance, relative to that found in
[25]
the parent helicate [Ln2(L2)3]6+
.
With this in mind, we
have kept a single average intermetallic interaction DELn,Ln
for macro(bi)cyclic complexes, but introduced a specific
Ln,Ln
single strand
values DE
for [Ln2(L12b-2H)]4+, which transforms
Equation (10) into Equation (17).
Ln;L12b
2;1
Ln;L12b
N2O
Ln;Ln
2
b
¼ 36ðf
Þ ðusingle strand
Þ
ð17Þ
along the series for L2 (10ꢀ7.3 m (Ln=La)ꢄceff ꢄ10ꢀ4.3
m
A second round of multilinear least-squares fits of Equa-
tions (8), (9), and (14)–(17) for Ln=Y, Lu obviously im-
proves the quality of the fit, but indeed confirms the coun-
(Ln=Lu); Table S16), which contrasts with its parallel de-
crease for L12b (10ꢀ6.8 m (Ln=La)ꢅceff ꢅ10ꢀ9.5 m (Ln=Lu);
Ln,L
intra
Table 3). This translates into an opposite trend for DG
=
terintuitive sequence DELn,Ln
single strand >DELn,Ln (Table 3, rows 9–
DG ꢀRTln(ceff) in Figure 6, and we can thus safely assign
Ln,L
intra
18; Figure S12 in the Supporting Information), which may
now rationalize the low stability constants observed for
[Ln2(L12b-2H)]4+ with large LnIII.
From this complete thermodynamic picture for the forma-
tion of [Ln2(L12b-2H)n](6ꢀ2n)+, it is tempting to assign the
the extremely low values of the effective concentration
(10ꢀ7ꢀ10ꢀ9 m) in L12b to the loss of two degrees of rotation-
al freedom, a factor that more severely affects the macrocyc-
lization processes around smaller lanthanides.
12728
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12719 – 12732

Binuclear Lanthanide Helicates
FULL PAPER
that the respective counterions, TfOꢀ/(nBu)4N+, have non-
negligible interactions with their respective partners, and
that their release into solution during the metal–ligand com-
plexation process limits the classical charge-neutralization
process described in water. This behavior may explain the
decrease in stability by twenty-five orders of magnitude in
going from [Ln2(L3-2H)3] (measured in water)[3] to
[Ln2(L12b-2H)3] (measured in CH2Cl2/CH3OH) or
[Ln2(L11-2H)3] (measured in CH3CN).[13] Beyond this cru-
cial factor, the comparison of the assembly of the negatively
charged segmental ligand [L12b-2H]2ꢀ with that of its neu-
tral analogue L2, performed in solvents of comparable po-
larities, shows that the effective concentration plays a lead-
ing role. Firstly, it decreases by approximately two orders of
magnitude in going from L2 to [L12b-2H]2ꢀ because of the
reduction of the total number of rotational degrees of free-
dom within the tridentate binding units that possess fused
hydroxyquinoline groups. This eventually disfavors the mac-
rocyclization processes that lead to [Ln2(L12b-2H)3]. Sec-
ondly, the effective concentration decreases along the lan-
thanide series for [L12b-2H]2ꢀ, whereas it increases for L2,
which produces opposite contributions to the quantities of
linear over macrocyclic complexes formed along the series.
This intimate dependence of the effective concentration on
minor changes in the number of rotational degrees of free-
dom within the tridentate binding unit opens novel possibili-
ties for rationally programming segmental ligands that favor
the thermodynamic formation of discrete polynuclear com-
plexes, or infinite coordination polymers,[40] with a predeter-
mined number of bridging ligands—a strong point for the
design of single-stranded lanthanide polymers extending
along one direction and working as organic light-emitting
diodes (OLEDs).[25] Finally, both DEL,L and DELn,Ln may
bring some additional help for the fine thermodynamic
tuning of the target assemblies. However, their physical
origin arising from a balance between opposite and huge
coulombic (repulsive) and solvation (attractive) interactions
strongly limits their intuitive chemical programming.[18,19] In
[Ln2(L12b-2H)3], the bowl-shaped curve found for DELn,Ln
is responsible for some variable contributions to the anti-co-
operative processes, which eventually provides an extra sta-
bilization for mid-range lanthanides, a remarkable behavior
that could be exploited for producing heterometallic f–f sys-
tems that show some planned deviations from statistics.[41]
Figure 6. Plots of four microscopic thermodynamic parameters in the
form of free energies of interactions, as a function of the inverse of nine-
coordinate ionic radii for the assemblies of a) [Ln(L10b-H)n](3ꢀn)+ and
[Ln2(L12b-2H)n](6ꢀ2n)+ (n=1–3, CH2Cl2/CH3OH 1:1, 298 K), and
b) [Ln(L13)n]3+ and [Ln2(L2)n]6+ (n=1–3, CH3CN, 298 K). The dotted
lines are only guides for the eyes.
Conclusion
Reconsidering the specifications mentioned in the introduc-
tion, we have indeed established the structure of the triple-
stranded binuclear helicates [Ln2(L12b-2H)3] and shown
that it is similar to those of [Ln2(L3-2H)3] and [Ln2(L2)3]6+
in the solid state and in solution, despite the fusion of the
terminal aromatic rings in L12b. On the other hand, the
thermodynamic assembly process is specific for each ligand
strand, and is strongly influenced by the choice of the sol-
vent. In weakly polar organic solvents, the free-energy
Experimental Section
General: Chemicals were purchased from Fluka AG and Aldrich and
used without further purification unless otherwise stated. Starting syn-
thons 4,[21] 5,[21] and ligands L2[2] and L13[25] were prepared according to
literature procedures. The syntheses of intermediates 1, 2, and 3 were in-
spired from the literature,[20] but the modifications and improvements
brought in this work justified a novel description of those three steps.
changes in the connection of Ln3+ to an N2O tridentate
Ln,L12b
inter
binding site are similar for [L12b-2H]2ꢀ or L2 (DG
=
Ln,L2
DG
ꢁꢀ42 kJmolꢀ1), and we deduce that the huge contri-
inter
bution of charge neutralization to the entropy of complexa-
tion (through the relaxation of the organization of the sol-
vent molecules),[16] is mainly restricted to water or to highly
polar solvents, in which the formation of ionic pairs is negli-
gible. For the Ln3+/[L12b-2H]2ꢀ system, we must consider
.
The trifluoromethanesulfonate salts Ln(CF3SO3)3 xH2O were prepared
from the corresponding oxide (Aldrich, 99.99%).[42] The Ln contents of
solid salts were determined by complexometric titrations with Titriplex
III (Merck) in the presence of urotropine and xylene orange.[43] Acetoni-
Chem. Eur. J. 2009, 15, 12719 – 12732
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12729

E. Terazzi, C. Piguet et al.
trile and dichloromethane were distilled over calcium hydride. Methanol
was distilled over Mg(OCH3)2. Silica gel plates Merck 60 F254 were used
for thin-layer chromatography (TLC) and Fluka silica gel 60 (0.04–
0.063 mm) was used for preparative column chromatography.
was added to the aqueous layer. A suspension was thus partitioned be-
tween a red aqueous phase (pH 8.0) and a yellow organic one. H2O2
(5 mL) was added and the pH was adjusted to 8.5 with an aqueous
NH4OH (25%) solution. The biphasic solution was stirred for 30 min and
the organic layer was separated from the dark-red aqueous phase, which
was further extracted with CH2Cl2 (3ꢄ100 mL). The combined organic
layers were dried over anhydrous MgSO4, filtered, and evaporated to
dryness. The white solid was purified by column chromatography (silica
gel, CH2Cl2/MeOH 98:2) to yield 7 (3.50 g, 4.54 mmol, 79%) as a white
solid. When required, 7 can be crystallized from CH3CN or by slow evap-
oration from a CH2Cl2/MeOH mixture. X-ray quality crystals can be ob-
tained with the latter method. 1H NMR (400 MHz, CDCl3): d=1.32 (t,
3J=7.1 Hz, 6H), 4.31 (s, 2H), 5.03 (q, 3J=7.1 Hz, 4H), 5.32 (s, 4H), 7.19
(dd, 3J=7.6 Hz, 4J=1.3 Hz, 2H), 7.26 (dd, 3J=8.4 Hz, 4J=1.5 Hz, 2H),
7.34–7.60 (m, 16H), 7.79 (s, 2H), 8.27 (d, 3J=8.5 Hz, 2H), 8.6 ppm (d,
3J=8.5 Hz, 2H); 13C NMR (100 MHz, CDCl3): d=15.20, 41.11, 70.82,
109.47, 109.98, 119.66, 120.08, 122.03, 124.99, 127.34, 128.10, 128.18,
128.51, 128.89, 132.24, 136.16, 136.37, 136.82, 139.51, 143.08, 149.19,
149.49, 155.08 ppm; ESIMS (CH2Cl2/MeOH 99:1): m/z: 771.5 [M+H]+,
1542.3 [2M+H]+.
Preparation of 1: 8-Hydroxyquinaldine (16.53 g, 103.84 mmol), benzyl
bromide
(26.64 g,
155.75 mmol),
anhydrous
K2CO3
(28.70 g,
207.66 mmol), and a catalytic amount of KI were suspended in acetone
(150 mL), and heated at reflux until complete consumption of the start-
ing materials (12 h, TLC: CH2Cl2/MeOH 98:2). The reaction mixture was
evaporated to dryness and the residue dissolved in CH2Cl2/H2O 1:1
(300 mL). The organic layer was separated and dried over anhydrous
MgSO4, filtered, and evaporated. The yellow oil was crystallized in
hexane (200 mL/08C), filtered, washed with cold hexane, and dried for
24 h under vacuum to give 1 (20.50 g, 82.23 mmol, 79%) as white crystals.
1H NMR (400 MHz, CDCl3): d=2.84 (s, 3H), 5.49 (s, 2H), 7.02 (dd, 3J=
7.6 Hz, 4J=1.5 Hz, 1H), 7.28–7.42 (m, 6H), 7.52–7.58 (m, 2H), 8.03 ppm
3
(d, J=8.4 Hz, 1H).
Preparation of 2: Compound 1 (15.00 g, 60.17 mmol) and SeO2 (8.34 g,
75.16 mmol) were suspended in dioxane (120 mL). The white solid quick-
ly dissolved and the color of the suspension turned red in a few minutes.
The reaction mixture was then heated at 808C under an inert atmos-
phere. After 2 h, the suspension turned brown and no more starting ma-
terial could be detected (TLC: CH2Cl2/MeOH 98:2). Selenium was fil-
tered through Celite, and washed with dichloromethane. The yellow fil-
trate was evaporated, dried under vacuum for 12 h, redissolved in CH2Cl2
(100 mL), and kept at RT for 24 h. The red selenium precipitate was re-
moved by filtration and the filtrate evaporated to dryness. The crude
product was purified by column chromatography (silica gel, CH2Cl2/
MeOH 100:0!99:1) to give 2 (13.05 g, 49.56 mmol, 83%) as a yellow-
brown solid. 1H NMR (400 MHz, CDCl3): d=5.52 (s, 2H), 7.18 (d, 3J=
7.8 Hz, 1H), 7.33–7.61 (m, 7H), 8.09 (d, 3J=8.3 Hz, 1H), 8.30 (d, 3J=
8.3 Hz, 1H), 10.35 ppm (s, 1H).
Preparation of L12b: The reaction was carried out under a nitrogen at-
mosphere. Compound 7 (1.50 g, 1.95 mmol) was dissolved in dry CH2Cl2
(50 mL). A BBr3 solution in CH2Cl2 (1m, 19.5 mL, 19.50 mmol) was
added through a septum. The colorless solution quickly turned red and a
beige precipitate formed, which was slowly transformed into a reddish
oil. The reaction mixture was stirred for 12 h and methanol (30 mL) was
slowly added to destroy the excess of BBr3 (a strongly exothermic reac-
tion). The reaction mixture was evaporated and dried under vacuum. The
red solid was suspended in H2O (100 mL) and sonicated for 5 min to
obtain an homogeneous suspension. Aqueous saturated NaHCO3 was
slowly added (over about 1 h) to adjust the pH to 7.5. The colorless pre-
cipitate was filtered and washed with water to completely remove benzyl
bromide. The solid was dried for 2 h under vacuum at 1208C and yielded
L12b (1.08 g, 1.83 mmol, 94%) as a beige solid. 1H NMR (400 MHz,
[D6]DMSO): d=1.40 (t, 3J=7.1 Hz, 6H), 4.25 (s, 2H), 5.07 (q, 3J=
7.1 Hz, 4H), 7.20 (dd, 3J=7.2 Hz, 4J=1.6 Hz, 2H), 7.31 (dd, 3J=8.5 Hz,
4J=1.6 Hz, 2H), 1.44–1.53 (m, 4H), 7.64 (d, 3J=8.5 Hz, 2H), 7.70 (s,
Preparation of 3: Compound 2 (10.00 g, 37.98 mmol) was dissolved in
THF (250 mL), H2O (175 mL) was added, and the pale yellow solution
was cooled at 08C. H2NSO3H (14.75 g, 151.92 mmol) and NaClO2
(13.74 g, 151.92 mmol) were added simultaneously and the color of the
stirred solution rapidly turned orange. The complete consumption of the
starting material required 30 min (TLC: CH2Cl2/MeOH 80:20). Water
(1.5 L) was slowly added under sonication to precipitate a white solid,
which was filtered, washed with water, and dried for 24 h under vacuum
3
3
2H), 8.41 (d, J=8.7 Hz, 2H), 8.45 (d, J=8.7 Hz, 2H), 9.83 ppm (s, 2H);
ESIMS (DMSO): m/z: 591.5 [M+H]+, 1182.3 [2M+H]+; elemental anal-
ysis calcd (%) for C37H30N6O2·1.17H2O: C 72.63, H 5.33, N 13.74; found:
C 72.67, H 5.04, N 13.41.
to give
3 (10.50 g, 37.88 mmol, 99%) as a
white solid. 1H NMR
(400 MHz, CDCl3): d=5.38 (s, 2H), 7.24 (dd, 3J=7.8 Hz, 4J=1.2 Hz,
1H), 7.37–7.65 (m, 7H), 8.32 (d, 3J=8.4 Hz, 1H), 8.40 ppm (d, 3J=
8.4 Hz, 1H).
Preparation of the [Ln2(L12b-2H)3] (Ln=La, Nd, Eu, Y, and Lu) com-
plexes: Ligand L12b (20 mg, 3.38ꢄ10ꢀ5 mol, 1 equiv) was suspended in
CH3CN (20 mL), shortly sonicated, and Ln(OTf)3·xH2O (2.27ꢄ10ꢀ5 mol)
was added. The solid rapidly disappeared and the color of the solution
changed from colorless to yellow or pale orange after 5 min heating at
658C. An excess of solid K2CO3 (2 equiv per phenol function) was added
and the solution was stirred at 658C for 12 h. During this time the solu-
tions turned orange. The Ln=La, Nd, and Eu solutions are limpid after
this treatment, whereas we observed the formation of an orange precipi-
tate for the Ln=Y- and Lu-containing solutions. The reaction mixtures
were precipitated with Et2O (50 mL) and cooled (08C). The orange pre-
cipitates were filtered and washed with water to remove all the inorganic
salts, then washed with Et2O and dried under vacuum at 1008C for 24 h
to give binuclear complexes [Ln2(L12b-2H)3]·xH2O (Table S4 in the Sup-
porting Information). The resulting solids varied from yellow [La2(L12b-
2H)3] to dark red [Lu2(L12b-2H)3]. Suitable X-ray quality crystals of
[Nd2(L12b-2H)3] were obtained by slow diffusion of Et2O into a diluted
solution of the complex in CH3CN.
Preparation of 6: The reaction was carried out under a nitrogen atmos-
phere. A large excess of compound 3 (5.00 g, 27.89 mmol) was dissolved
in dry CH2Cl2 (100 mL). Then, SOCl2 (21.30 g, 179.04 mmol) and a cata-
lytic amount of DMF (50 mL) were added. The solution changed from
colorless to orange and a precipitate was formed. The reaction mixture
was heated at reflux until the precipitate disappeared (1 h). The resulting
solution was evaporated to dryness and the yellow solid was dried under
vacuum at 808C for 1 h. This activated intermediate was dissolved in dry
CH2Cl2 (75 mL), and compound 5 (2.06 g, 5.97 mmol) was added as a
solid. The solution was heated at reflux for 1 h, at which point no starting
material remained (TLC: CH2Cl2/MeOH 98:2). The pH of the reaction
was adjusted to 7 by addition of N,N’-diisopropylethylamine. The reac-
tion mixture was evaporated and dried under vacuum. The voluminous
yellow solid was purified by column chromatography (silica gel, CH2Cl2/
MeOH 100:0 !98:2) to give 6 (5.18 g, 5.97 mmol, 100%) as a yellow
solid. ESIMS (CH2Cl2/MeOH 99:1): m/z: 867.9 [M+H]+.
Preparation of [Ln2(L12b-2H)3] (Ln=La, Nd, Eu, and Y) samples for
1H NMR spectroscopy in CD3CN: Ligand L12b (5.0 mg, 8.46ꢄ
10ꢀ6 mmol, 1 equiv) was suspended in CD3CN (0.5 mL) and shortly soni-
cated. Ln(OTf)3·xH2O (5.67ꢄ10ꢀ6 mmol, 0.67 equiv) was dissolved in
CD3CN (0.25 mL). The lanthanide solution was poured into the ligand
solution. The volume of the reaction mixture was adjusted to 1 mL by ad-
dition of CD3CN (0.25 mL). The mixture was heated at 658C for 30 min
to get a yellow or pale orange solution. An excess of solid K2CO3 (about
2 equiv per phenol function) was added and the suspension was stirred at
Preparation of 7: Compound 6 (5.00 g, 5.77 mmol) and activated iron
powder (6.44 g, 115 mmol) were suspended in EtOH/H2O 2:1 (300 mL).
The reaction mixture was heated to 608C to solubilize 6, and HCl (37%)
was slowly added until the reaction started to bubble (15 mL). The reac-
tion mixture was then heated at reflux for 12 h. The excess iron powder
was removed by decantation, washed with dichloromethane, and the or-
ganic solvents were evaporated. A solution of Na4EDTA (60 g; EDTA=
ethylenediaminetetraacetic acid) in H2O (300 mL) and CH2Cl2 (500 mL)
12730
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12719 – 12732

Binuclear Lanthanide Helicates
FULL PAPER
658C for 1 h. Filtration through a membrane (0.25 mm porosity) directly
in a 5 mm NMR spectroscopy tube allowed the subsequent recording of
the 1H NMR spectra. An orange precipitate formed upon waiting, a ten-
dency that was more pronounced for small lanthanide ions.
ities A-[Nd2(L12b-2H)3] (multiplicity: 8) and B-[Nd2(L12b-2H)3] (multi-
plicity: 4) were found in the unit cell together with interstitial solvent
molecules, thus leading to the complete chemical formula {A-[Nd2(L12b-
2H)3]}2·{B-[Nd2(L12b-2H)3]}·8CH3CN·4H2O (10) with Z=4. B-
[Nd2(L12b-2H)3] was located on a crystallographic twofold axis passing
Preparation of the samples for 1H NMR spectroscopy and ESIMS titra-
tions in [D6]DMSO: NBu4OH·30H2O (0.5 mL, 1.00ꢄ10ꢀ2 m, 2 equiv) in
through C19e. All hydrogen atoms were calculated and fixed (Uiso
=
0.06 ꢀ2). All other atoms were refined with isotropic atomic displace-
ment parameters, except the Nd atoms (anisotropic). The ethyl groups
connected to the benzimidazole rings were disordered and gave large Uiso
parameters, and so did the terminal hydroxyquinoline rings, which were
thus refined with restraints on bond lengths and bond angles (235 re-
straints). The solvent molecules were refined with fixed Uiso and blocked
during the refinement except for one acetonitrile molecule (N2sC3sC4s).
According to the limited quality of this crystal structure, but considering
its importance for confirming the formation of a triple-stranded helicate
with intermetallic contact distance of 9.0 ꢀ, we have not incorporated it
into the CCDC file, but have instead provided the cif file in the Support-
ing Information.
[D6]DMSO was added to
a
solution of L12b (5.00ꢄ10ꢀ3 m) in
[D6]DMSO. The solution immediately turned red and the 1H NMR spec-
trum confirmed the quantitative formation of [L12b-2H](nBu4N)2. After
a few minutes, the solution turned brown and side products precipitated.
Consequently, aliquots of Ln(OTf)3·xH2O (Ln=La, Nd, Eu, Y, and Lu;
0.67, 1.0, and 2.0 equiv) from stock solutions (3.35ꢄ10ꢀ3 m in [D6]DMSO)
were to be added immediately. The metal-stabilized solutions were then
stirred at 658C for 1 h (precipitates disappeared after 5 min) before re-
1
cording H NMR spectra at 258C. The solutions were prepared under the
same conditions for ESIMS titrations, except that 1) nondeuterated
DMSO was used, and 2) the solutions were eventually diluted 10 times
and 2% MeOH was added before recording the spectra.
Spectroscopic measurements: Spectrophotometric titrations were per-
formed using a J&M diode array spectrometer (Tidas series) connected
to an external computer. In a typical experiment, 50 mL of ligand in
CH2Cl2/MeOH 1:1 (10ꢀ4 m) were titrated at 298 K with a solution of
Ln(OTf)3·xH2O (10ꢀ3 m) in CH2Cl2/MeOH 1:1 under an inert atmos-
phere. After each addition of 0.20 mL, the absorbance was recorded
using Hellma optrodes (optical path length 0.1 cm) immersed in the titra-
tion vessel controlled by a thermostat and connected to the spectrometer.
Mathematical treatment of the spectrophotometric titrations was per-
formed using factor analysis[32] and with the SPECFIT program.[33] 1H and
Acknowledgements
We thank Johan Varin and Laetitia Gaillard for technical support. Finan-
cial support from the Swiss National Science Foundation is gratefully ac-
knowledged.
13C NMR spectra were recorded at 298 K using
a Bruker Avance
[1] a) C. Piguet, G. Bernardinelli, A. F. Williams, Angew. Chem. 1992,
102, 1530–1532; Angew. Chem. Int. Ed. Engl. 1992, 31, 1622–1624;
b) C. Piguet, J.-C. G. Bꢃnzli, G. Bernardinelli, G. Hopfgartner, A. F.
Williams, J. Am. Chem. Soc. 1993, 115, 8197–8296.
[2] a) N. Martin, J.-C. G. Bꢃnzli, V. McKee, C. Piguet, G. Hopfgartner,
Inorg. Chem. 1998, 37, 577–589; b) K. Zeckert, J. Hamacek, J.-P.
Rivera, S. Floquet, A. Pinto, M. Borkovec, C. Piguet, J. Am. Chem.
Soc. 2004, 126, 11589–11601.
400 MHz spectrometer. Chemical shifts are given in ppm with respect to
TMS. Pneumatically assisted electrospray mass spectra (ESIMS) were re-
corded from 10ꢀ4 m solutions using an Applied Biosystems API 150EX
LC/MS system equipped with a Turbo Ionspray source. Least-square fits
were performed using Microsoft Excel.[44] Elemental analyses were per-
formed by Dr. H. Eder from the Microchemical Laboratory of the Uni-
versity of Geneva.
[3] a) M. Elhabiri, R. Scopelliti, J.-C. G. Bꢃnzli, C. Piguet, J. Am. Chem.
Soc. 1999, 121, 10747–10762; b) M. Elhabiri, J. Hamacek, J.-C. G.
Bꢃnzli, A. M. Albrecht-Gary, Eur. J. Inorg. Chem. 2004, 51–62.
[4] a) S. Floquet, N. Ouali, B. Bocquet, G. Bernardinelli, D. Imbert, J.-
C. G. Bꢃnzli, G. Hopfgartner, C. Piguet, Chem. Eur. J. 2003, 9,
1860–1875; b) S. Floquet, M. Borkovec, G. Bernardinelli, A. Pinto,
L.-A. Leuthold, G. Hopfgartner, D. Imbert, J.-C. G. Bꢃnzli, C.
Piguet, Chem. Eur. J. 2004, 10, 1091–1105.
[5] a) K. Zeckert, J. Hamacek, J.-M. Senegas, N. Dalla-Favera, S. Flo-
quet, G. Bernardinelli, C. Piguet, Angew. Chem. 2005, 117, 8168–
8172; Angew. Chem. Int. Ed. 2005, 44, 7954–7958; b) N. Dalla-
Favera, J. Hamacek, M. Borkovec, D. Jeannerat, F. Gumy, J.-C. G.
Bꢃnzli, G. Ercolani, C. Piguet, Chem. Eur. J. 2008, 14, 2994–3005.
[6] J. L. Lessmann, W. de W. Horrocks, Jr., Inorg. Chem. 2000, 39,
3114–3124.
X-ray crystallography: Summary of crystal data, intensity measurements,
and structure refinements for 7, and A-[Nd2(L12b-2H)3]}2·{B-[Nd2(L12b-
2H)3]}·8CH3CN·4H2O (10) are collected in Table S17 in the Supporting
Information. All crystals were mounted on quartz fibers with protection
oil. Cell dimensions and intensities were measured at 150 K using a Stoe
IPDS diffractometer with graphite-monochromated MoKa radiation (l=
0.71073 ꢀ). Data were corrected for Lorentz and polarization effects and
for absorption. The structures were solved by direct methods (SIR97),[45]
and all other calculation were performed using ShelX97 (7)[46] XTAL
(10)[47] systems and ORTEP[48] programs. CCDC-739143 (7) contains the
supplementary crystallographic data for this paper. These data can be ob-
tained free of charge from The Cambridge Crystallographic Data Centre
via www.ccdc.cam.ac.uk/data_request/cif. The cif file of 10 is given in the
Supporting Information.
Comments on the crystal structure of 7: All non-hydrogen atoms (59)
were refined with anisotropic atomic displacement parameters. The hy-
drogen atoms were observed and refined with Uiso =0.06 ꢀ2, except for
one methyl (H40 to H42) and those of one terminal phenyl group (H8 to
H12), for which the atomic positions of the hydrogen atoms were calcu-
lated and fixed during the refinement. The phenyl C11–C16 atoms were
slightly disordered, but attempts to split the carbon atoms for forming
two phenyl rings with variable population parameters did not improve re-
finement nor convergence. The maximum residual electronic density was
located close to this aromatic cycle.
[7] a) C. D. B. Vandevyver, A.-S. Chauvin, S. Comby, J.-C. G. Bꢃnzli,
Chem. Commun. 2007, 1716–1718; b) A.-S. Chauvin, S. Comby, B.
Song, C. D. B. Vandevyver, F. Thomas, J.-C. G. Bꢃnzli, Chem. Eur. J.
2007, 13, 9515–9526; c) A.-S. Chauvin, S. Comby, B. Song, C. D. B.
Vandevyver, J.-C. G. Bꢃnzli, Chem. Eur. J. 2008, 14, 1726–1739;
d) E. Deiters, B. Song, A.-S. Chauvin, C. D. B. Vandevyver, J.-C. G.
Bꢃnzli, New J. Chem. 2008, 32, 1140–1152; e) J.-C. G. Bꢃnzli, A.-S.
Chauvin, C. D. B. Vandevyver, S. Bo, S. Comby, Ann. N. Y. Acad.
Sci. 2008, 1130, 97–105; f) B. Song, C. D. B. Vandevyver, A.-S.
Chauvin, J.-C. G. Bꢃnzli, Org. Biomol. Chem. 2008, 6, 4125–4133;
g) B. Song, C. D. B. Vandevyver, E. Deiters, A.-S. Chauvin, I. Hem-
milꢅ, J.-C. G. Bꢃnzli, Analyst 2008, 133, 1749–1756; h) J.-C. G.
Bꢃnzli, Chem. Lett. 2009, 38, 104–109; i) E. Deiters, B. Song, A.-S.
Chauvin, C. D. B. Vandevyver, F. Gumy, J.-C. G. Bꢃnzli, Chem. Eur.
J. 2009, 15, 885–900.
Comments on the crystal structure of 10: More than two hundred crystal-
lization attempts provided some extremely fragile red crystals, from
which only data collections of limited qualities could be obtained (high-
angle reflections with small intensities, despite long exposition periods of
10 min, and attempts to increase the imaging plate (IP)/crystal distance
of 90 mm to improve resolution at the cost of limiting the angular
domain of 2qmax =458). In the best data collection, two independent neu-
tral binuclear triple-stranded helical complexes possessing opposite helic-
[8] a) A. Dossing, Eur. J. Inorg. Chem. 2005, 1425–1434; b) S. Comby,
J.-C. G. Bꢃnzli in Handbook on the Physics and Chemistry of Rare
Earths, Vol. 37 (Eds.: K. A. Gschneidner, Jr., J.-C. G. Bꢃnzli, V. K.
Chem. Eur. J. 2009, 15, 12719 – 12732
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12731

E. Terazzi, C. Piguet et al.
Pecharsky), Elsevier, Amsterdam, 2007, pp. 217–470; c) T. Nishioka,
K. Fukui, K. Matsumoto in Handbook on the Physics and Chemistry
of Rare Earths, Vol. 37 (Eds.: K. A. Gschneidner, Jr., J.-C. G. Bꢃnzli,
V. K. Pecharsky), Elsevier, Amsterdam, 2007, pp. 171–220; d) M. D.
Ward, Coord. Chem. Rev. 2007, 251, 1663–1677.
Bartczak, Acta Crystallogr. Sect. B 2005, 61, 429–434; c) F. Zocchi,
J. Mol. Struct. 2007, 825–846, 73–78.
[29] A. Escande, L. Guꢁnꢁe, K.-L. Buchwalder, C. Piguet, Inorg. Chem.
2009, 48, 1132–1147.
[30] I. D. Brown, Acta Crystallogr. Sect. B 1977, 33, 1305–1310.
[31] a) E. Leize, A. Jaffrezic, A. van Dorsselaer, J. Mass Spectrom. 1996,
31, 537–544; b) G. Hopfgartner, F. Vilbois, C. Piguet, Rapid
Commun. Mass Spectrom. 1999, 13, 302–306.
[32] E. R. Malinowski, D. G. Howery, Factor Analysis in Chemistry,
Wiley, New York, 1980.
[33] a) H. Gampp, M. Maeder, C. J. Meyer, A. Zuberbꢃhler, Talanta
1985, 32, 1133–1139; b) H. Gampp, M. Maeder, C. J. Meyer, A. Zu-
berbꢃhler, Talanta 1986, 33, 943–951.
[34] a) C. Piguet, G. Bernardinelli, G. Hopfgartner, Chem. Rev. 1997, 97,
2005–2062; b) M. Albrecht, Chem. Rev. 2001, 101, 3457–3497.
[35] The diastereotopic methylene protons of the ethyl residues (H10 in
Figure 5) provide two pseudo-sextets (ABX3 spin system with J2
ꢁ2J3) without implying any blocked rotations as suggested in
ref. [12].
[9] A. Nonat, D. Imbert, J. Pꢁcaut, M. Giraud, M. Mazzanti, Inorg.
Chem. 2009, 48, 4207–4218.
[10] a) D. Imbert, S. Comby, A.-S. Chauvin, J.-C. G. Bꢃnzli, Chem.
Commun. 2005, 1432–1434; b) M. Albrecht, O. Osetska, J. Klanker-
mayer, R. Frçhlich, F. Gumy, J.-C. G. Bꢃnzli, Chem. Commun. 2007,
1834–1836; c) M. Albrecht, M. Fiege, O. Osetska, Coord. Chem.
Rev. 2008, 252, 812–824.
[11] M. Albrecht, O. Osetska, R. Frçhlich, Dalton Trans. 2005, 3757–
3762.
[12] N. M. Shavaleev, R. Scopelliti, F. Gumy, J.-C. G. Bꢃnzli, Inorg.
Chem. 2008, 47, 9055–9068.
[13] M. Albrecht, O. Osetska, R. Frçhlich, J.-C. G. Bꢃnzli, A. Aebischer,
F. Gumy, J. Hamacek, J. Am. Chem. Soc. 2007, 129, 14178–14179.
[14] G. Ercolani, C. Piguet, M. Borkovec, J. Hamacek, J. Phys. Chem. B
2007, 111, 12195–12203.
[15] J. Hamacek, M. Borkovec, C. Piguet, Dalton Trans. 2006, 1473–
1490, and references therein.
[36] K. Binnemans, K. Lodewyckx, T. Cardinaels, T. N. Parac-Vogt, C.
Bourgogne, D. Guillon, B. Donnio, Eur. J. Inorg. Chem. 2006, 150–
157.
[16] a) R. J. Motekaitis, A. E. Martell, R. A. Hancock, Coord. Chem.
Rev. 1994, 133, 39–65; b) P. Caravan, T. Hedlund, S. Liu, S. Sjçberg,
C. Orvig, J. Am. Chem. Soc. 1995, 117, 11230–11238; c) J.-M. Sene-
gas, G. Bernardinelli, D. Imbert, J.-C. G. Bꢃnzli, P.-Y. Morgantini, J.
Weber, C. Piguet, Inorg. Chem. 2003, 42, 4680–4695.
[17] N. Dalla-Favera, J. Hamacek, M. Borkovec, D. Jeannerat, F. Gumy,
J.-C. G. Bꢃnzli, G. Ercolani, C. Piguet, Chem. Eur. J. 2008, 14, 2994–
3005.
[18] G. Canard, C. Piguet, Inorg. Chem. 2007, 46, 3511–3522.
[19] T. Riis-Johannessen, N. Dalla Favera, T. Todorova, S. M. Huber, L.
Gagliardi, C. Piguet, Chem. Eur. J. 2009, DOI: 10.1002/
chem.200900904.
[20] C. Caris, P. Baret, J.-L. Pierre, G. Serratrice, Tetrahedron 1996, 52,
4659–4672.
[21] C. Piguet, G. Bernardinelli, B. Bocquet, A. Quattropani, A. F. Wil-
liams, J. Am. Chem. Soc. 1992, 114, 7440–7491.
[22] C. Piguet, B. Bocquet, G. Hopfgartner, Helv. Chim. Acta 1994, 77,
931–942.
[23] F. H. Allen, O. Kennard, D. G. Watson, L. Brammer, O. A. G. R.
Taylor, J. Chem. Soc. Perkin Trans. 2 1987, S1–S19.
[37] a) S. W. Benson, J. Am. Chem. Soc. 1958, 80, 5151–5154; b) S. W.
Benson, Thermochemical Kinetics, 2nd ed., Wiley-Interscience, New
York, 1976, pp. 37–39.
[38] M. R. Willcott, R. E. Lenkinski, R. E. Davis, J. Am. Chem. Soc.
1972, 94, 1742–1744,
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
P
P
Ln,L
i,exp
Ln,L
i,calcd
Ln,L
2
2
AFLn
½
ðlog b ꢀlog b
Þ ꢂ=½
ðlog bi,expÞ ꢂ.
i
i
[39] A. Mulder, J. Huskens, D. N. Reinhoudt, Org. Biomol. Chem. 2004,
2, 3409–3424. The effective concentration (ceff) is often referred to
as the effective molarity (EM) when focusing on its experimental
value. For the sake of clarity, we will only use in this work the effec-
tive concentration for both the concept and its experimental value.
[40] C. Marchal, Y. Filinchuk, X.-Y. Chen, D. Imbert, M. Mazzanti,
Chem. Eur. J. 2009, 15, 5273–5288.
[41] N. Kerbellec, D. Kustaryono, V. Haquin, M. Etienne, C. Daigue-
bonne, O. Guillou, Inorg. Chem. 2009, 48, 2837–2843.
[42] J. F. Desreux in Lanthanide Probes in Life, Chemical and Earth Sci-
ences (Eds.: J.-C. G. Bꢃnzli, G. R. Choppin), Elsevier, Amsterdam,
1989, Chapter 2, p. 43.
[43] G. Schwarzenbach, Complexometric Titrations, Chapman & Hall,
London, 1957, p. 8.
[44] Microsoft Excel 2003, Microsoft Corporation, USA.
[45] A. Altomare, M. C. Burla, M. Camalli, G. Cascarano, C. Giacovaz-
zo, A. Guagliardi, G. Moliterni, G. Polidori and R. Spagna, J. Appl.
Crystallogr. 1999, 32, 115–119.
[46] SHELXL97, Program for the Solution and Refinement of Crystal
Structures, G. M. Sheldrick, University of Gçttingen, Gçttingen,
1997.
[47] XTAL 3.2, Userꢆs Manual (Eds.: S. R. Hall, H. D. Flack, J. M. Stew-
art), Universities of Western Australia and Maryland, 1989.
[48] ORTEP II; Report ORNL-5138, C. K. Johnson, Oak Ridge National
Laboratory: Oak Ridge, 1976.
[24] J. H. Brewster, Top. Curr. Chem. 1974, 46, 29–71.
[25] N. Dalla Favera, L. Guꢁnꢁe, G. Bernardinelli, C. Piguet, Dalton
Trans. 2009, DOI: 10.1039/b905131g.
[26] R. D. Shannon, Acta. Crystallogr. Sect. A 1976, 32, 751–767. The
ionic radii are calculated by using r(N)=1.46 ꢀ and r(O)=1.35 ꢀ
CN¼9
CN¼9
CN¼9
and amount to R
=1.17 ꢀ, R
=1.19 ꢀ and R
=1.16 ꢀ in
Nd1
Nd2
Nd3
CN¼9
10, and R
=1.17 ꢀ in HHT-[Nd(L10-2H)3].
Nd
[27] a) I. D. Brown, D. Altermatt, Acta Crystallogr. Sect. B 1985, 41,
244–247; b) N. E. Brese, M. OꢆKeeffe, Acta Crystallogr. Sect. B
1991, 47, 192–197; c) I. D. Brown, Acta Crystallogr. Sect. B 1992, 48,
553–572; d) I. D. Brown, The Chemical Bond in Inorganic Chemis-
try, Oxford University Press, Oxford, 2002.
[28] a) A. Trzesowska, R. Kruszynski, T. J. Bartczak, Acta Crystallogr.
Sect. B 2004, 60, 174–178; b) A. Trzesowska, R. Kruszynski, T. J.
Received: July 21, 2009
Published online: September 25, 2009
12732
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12719 – 12732