Smaller than a nanoparticle with the design of discrete polynuclear molecular complexes displaying near-infrared to visible upconversion

Article abstract of DOI:10.1039/c4dt02336f

This work shows that the operation of near-infrared to visible light-upconversion in a discrete molecule is not limited to non-linear optical processes, but may result from superexcitation processes using linear optics. The design of nine-coordinate metallic sites made up of neutral N-heterocyclic donor atoms in kinetically inert dinuclear [GaEr(L1)₃]⁶⁺ and trinuclear [GaErGa(L2)₃]⁹⁺ helicates leads to [ErN₉] chromophores displaying unprecedented dual visible nanosecond Er(⁴S_3/2→⁴I_15/2) and near-infrared microsecond Er(⁴I_13/2→⁴I_15/2) emissive components. Attempts to induce one ion excited-state absorption (ESA) upconversion upon near-infrared excitation of these complexes failed because of the too-faint Er-centred absorption cross sections. The replacement of the trivalent gallium cation with a photophysically-tailored pseudo-octahedral [CrN₆] chromophore working as a sensitizer for trivalent erbium in [CrEr(L1)₃]⁶⁺ improves the near-infrared excitation efficiency, leading to the observation of a weak energy transfer upconversion (ETU). The connection of a second sensitizer in [CrErCr(L2)₃]⁹⁺ generates a novel mechanism for upconversion, in which the superexcitation process is based on the Cr^III-sensitizers. Two successive Cr→Er energy transfer processes (concerted-ETU) compete with a standard Er-centred ETU, and a gain in upconverted luminescence by a factor larger than statistical values is predicted and observed.

Full text of DOI:10.1039/c4dt02336f

Dalton
Transactions
View Article Online
View Journal
PAPER
Smaller than a nanoparticle with the design
of discrete polynuclear molecular complexes
Cite this: DOI: 10.1039/c4dt02336f
displaying near-infrared to visible upconversion†‡
Davood Zare,a Yan Suffren,b Laure Guénée,c Svetlana V. Eliseeva,d
Homayoun Nozary,a Lilit Aboshyan-Sorgho,a Stéphane Petoud,*d Andreas Hauser*b
and Claude Piguet*a
This work shows that the operation of near-infrared to visible light-upconversion in a discrete molecule is
not limited to non-linear optical processes, but may result from superexcitation processes using linear
optics. The design of nine-coordinate metallic sites made up of neutral N-heterocyclic donor atoms in
kinetically inert dinuclear [GaEr(L1)3]6+ and trinuclear [GaErGa(L2)3]9+ helicates leads to [ErN9] chromo-
phores displaying unprecedented dual visible nanosecond Er(4S3/2→4I15/2) and near-infrared microsecond
Er(4I13/2→4I15/2) emissive components. Attempts to induce one ion excited-state absorption (ESA) up-
conversion upon near-infrared excitation of these complexes failed because of the too-faint Er-centred
absorption cross sections. The replacement of the trivalent gallium cation with a photophysically-tailored
pseudo-octahedral [CrN6] chromophore working as a sensitizer for trivalent erbium in [CrEr(L1)3]6+
improves the near-infrared excitation efficiency, leading to the observation of a weak energy transfer up-
conversion (ETU). The connection of a second sensitizer in [CrErCr(L2)3]9+ generates a novel mechanism
for upconversion, in which the superexcitation process is based on the CrIII-sensitizers. Two successive
Cr→Er energy transfer processes (concerted-ETU) compete with a standard Er-centred ETU, and a gain in
upconverted luminescence by a factor larger than statistical values is predicted and observed.
Received 1st August 2014,
Accepted 30th September 2014
DOI: 10.1039/c4dt02336f
www.rsc.org/dalton
changes), near-infrared (NIR) to visible light-upconversion pro-
cesses may appear puzzling. However, it has been known for a
Introduction
Compared to the classical spontaneous transformation of long time that the intense electromagnetic irradiation (typi-
energy in physics (energy degradation) and in chemistry (spon- cally in the range of MW cm−2) of polarizable materials may
taneous chemical reactions induced by negative free energy induce weak non-linear responses, which ultimately produce
photons of higher energy than those used for excitation.1
Despite the low quantum yields of such non-linear processes,
aDepartment of Inorganic and Analytical Chemistry, University of Geneva, 30 quai
E. Ansermet, CH-1211 Geneva 4, Switzerland. E-mail: Claude.Piguet@unige.ch
bDepartment of Physical Chemistry, University of Geneva, 30 quai E. Ansermet,
CH-1211 Geneva 4, Switzerland. E-mail: Andreas.Hauser@unige.ch
cLaboratory of Crystallography, University of Geneva, 24 quai E. Ansermet, CH-1211
Geneva 4, Switzerland
the transparency of living tissues to incident NIR light was
attractive enough to exploit these phenomena for the in vivo
sensitization of optical probes and sensors in photobiology.2
For technological applications, such huge input intensities are
of limited interest, and efforts are directed towards the design
of upconverting materials based on linear optical processes
and compatible with common light sources, for instance for
the conversion of the NIR part of the solar light into green
emission compatible with its absorption by dye-sensitized or
by crystalline silica solar cells (the power-density of the terres-
dCentre de Biophysique Moléculaire, CNRS UPR 4301, Rue Charles Sadron, F-45071
Orléans Cedex 2, France. E-mail: Stephane.Petoud@inserm.fr
†Submitted to Dalton Transactions dedicated themed issue: Advancing the chem-
istry of the f-elements.
‡Electronic supplementary information (ESI) available: Synthesis of ligand L1
(Appendix 1), determination of thermodynamic exchange constants (Appendix
2), kinetic analysis (Appendix 3) and calculation of normalized steady-state trial solar irradiance is approximately 0.1 W cm−2).3
population densities (Appendix 4). Tables collecting elemental analysis
In this context, the non-coherently sensitized upconversion
(Table S1), ESI-MS peaks (Tables S2–S3) and crystallographic data (Tables S4–S5).
based on triplet–triplet annihilation (TTA) is currently one of
Figures showing NMR spectra (Fig. S1–S4), ORTEP views (Fig. S5), packing inter-
the most efficient strategies.4 According to the mechanism
actions (Fig. S6–S8) and photophysical data (Fig. S9–S25). CCDC 1003567 and
depicted in Scheme 1, a sensitizer (S)/acceptor (A) pair of mole-
cules is required. The sensitizer S is responsible for the
1003568. For ESI and crystallographic data in CIF or other electronic format see
DOI: 10.1039/c4dt02336f
This journal is © The Royal Society of Chemistry 2014
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Scheme 1 Qualitative Jablonski diagrams illustrating the sensitized
triplet–triplet annihilation (TTA) upconversion process operating in a S/A
mixture. Solid arrows indicate transitions in which a photon is involved,
while black dashed arrows indicate radiationless processes (ISC = inter-
system crossing and TTET = triplet–triplet energy transfer). The alternat-
ing dashed-dotted red arrows stand for the triplet–triplet annihilation
mechanism.
Scheme 2 Kinetic schemes depicting the modelling of activator-
centred superexcitation processes according to (a) a one ion excited
state absorption (ESA) process and (b) an energy transfer upconversion
(ETU) process occurring upon off-resonance excitations. keAxc(0→1) and
keSxc(0→1) are the pumping rate constants for the irradiation into the acti-
1S→1S* light-harvesting process which is followed by an inter- vator, respectively sensitizer absorption bands, kAi→j and kiS→j are the
decay rate constants (i.e. the sum of radiative and non-radiative pro-
cesses) of level i into level j centred on the sensitizer and on the activa-
tor, respectively. WSn→A are the rate constants of sensitizer-to-activator
energy transfer processes. Red solid arrows correspond to excitation
photons and green solid arrows stand for upconverted emission.
system crossing (ISC) process leading to the establishment
of a long-lived triplet excited state 3S*. The subsequent
3S*/1A→1S/3A* Dexter-type triplet–triplet energy transfer pro-
vides an acceptor-centred triplet state 3A*, which is sufficiently
long-lived to diffuse and to collide with a second partner. The
triplet–triplet annihilation then produces a mixture of singlet,
triplet and quintet excited dimers, in which the 3A* +
upconversion could be detected upon very intense irradiation
3A*→1A* + 1A pathway leads to the targeted high-energy singlet
excited state on the acceptor. Relaxation of the A* state to the
of dimethylsulfoxide solutions containing Tm3+.8 Indirect exci-
tation using sensitizers for collecting the initial NIR irradiation
followed by energy transfer onto the acceptor (ETU,
Scheme 2b) may benefit from the larger absorption cross sec-
tions and longer lifetimes of the sensitizers,5 but to date only
little attention has been given to the implementation of ETU
in molecules. In 2003, Hoshino and coworkers reported a sur-
prising linear Er-centred upconversion process induced by the
intense irradiation of the low-energy tail of the ligand-absorp-
tion band in films of [Er(quinolinate)3] complexes, but its
alternative assignment to a ligand-centred non-linear optical
two-photon absorption followed by Er-centred luminescence
could not be excluded.9 Inspired by this pioneering work,
Piguet and co-workers used trivalent chromium for the indir-
ect sensitization of ErIII in the trinuclear triple-stranded
[CrErCr(L2)]9+ helicate.10 Continuous-wave NIR irradiation at
750 nm indeed generated the excitation of the chromium
cations via their spin-flip Cr(2E←4A2) and Cr(2T1←4A2) tran-
sitions, which were responsible for a weak, but reproducible
green upconverted erbium-centred Er(4S3/2→4I15/2) lumines-
cence at 540 nm (Scheme 3).10 The quadratic dependence of
the upconverted emission upon the incident intensity
observed in diluted solutions combined with the crucial role
played by Cr→Er energy transfer processes unambiguously
support the presence of an ETU mechanism operating within a
single supramolecular complex.11
1
ground state is finally accompanied by the emission of a
photon of higher energy than those involved in the excitation
process. Typical green-to-blue upconversion is experimentally
observed by using sensitizers and activators containing aro-
matic ring systems of various sizes. However, the requirements
for the diffusion and collision of two excited triplet acceptors
for TTA limit this methodology to intermolecular processes
occurring in solution, in rubbery polymeric materials or in
solid matrices, ensuring sufficiently efficient diffusion of mole-
cules and/or molecular excitons under anaerobic conditions
(since dioxygen can easily quench triplet excited states).4
Though less efficient in terms of quantum yields, the super-
excitation mechanisms depicted in Scheme 2 also rely on
linear and sequential optical processes. They are currently
implemented in low-phonon nanoparticles,5 but their engin-
eering can be theoretically further miniaturized and confined
in a single molecular entity. For the one ion excited state
absorption mechanism (ESA, Scheme 2a), the initial excitation
produces an intermediate excited state, the lifetime of which
must be long enough to allow for a second absorption process
to occur. Such a situation is difficult to implement in mole-
cules where the existence of high-energy oscillators drastically
reduces the excited-state lifetimes.6 Indeed, attempts to detect
ESA-processes in coordination complexes using trivalent
lanthanides as activators (Ln = Er, Tm, Yb) have failed to date,7
except for a very recent report claiming that NIR to visible ESA-
Further optimization requires a deeper level of understand-
ing of the ETU mechanism in [CrErCr(L2)]9+. This work can be
Dalton Trans.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
vents luminescence in coordination complexes, except for
the fluorescence of the lowest Er(4I13/2) excited state for which
p ≥ 3.7
Efforts made for protecting the erbium-activator from high-
energy C–H oscillators in coordination complexes succeeded
in extending the Er(4I13/2) excited state lifetimes by more than
one order of magnitude,13i and we therefore planned to take
advantage of this strategy with the introduction of rigid un-
saturated bis-(N-methyl-benzimidazol-2-yl)pyridine tridentate
units L0 into the segmental ligands L1 and L2 (Scheme 4, see
Appendix 1 in the ESI‡). Reaction of L0 with Ln(CF3SO3)3 was
known to give poorly stable triple-helical [Ln(L0)3]3+ complexes
with the small Y3+ and Er3+ cations,14 but the formation of
these [LnN9] chromophores was favoured by the self-assembly
of L1 and L2 with Ln(CF3SO3)3 in the presence of Ga(CF3SO3)3.
The heterometallic dinuclear [GaLn(L1)3](CF3SO3)6·mC2H5CN·
nH2O and trinuclear [GaLnGa(L2)3](CF3SO3)9·mC2H5CN·nH2O
Scheme 3 Jablonski diagrams obtained from absorption and emission
spectra recorded for the different chromophores in [CrErCr(L2)3]-
(CF3SO3)9. Black dashed arrows indicate radiationless sensitizer-to-acti-
vator energy transfer processes and curled arrows stand for non-radia-
tive internal conversion. Red solid arrows correspond to the NIR
excitation photons and the green solid arrow stands for the visible
upconverted emission. Adapted from ref. 11.
achieved thanks to the preparation of analogous complexes, in
which (i) the photophysically active CrIII and ErIII chromo-
phores are replaced by ‘innocent’ partners and (ii) the number
of Cr-sensitizers per Er-activators is increased stepwise from
0 to 2. Here we thus report on the chemical strategy used for
the design of isostructural polynuclear heterometallic gallium–
lanthanide and chromium–lanthanide complexes, in which
energy transfer upconversion can be deciphered and
optimized.
Results and discussion
Optimizing the erbium coordination sphere for its use as an
activator in molecular upconversion processes
Except for the report of Hoshino and co-workers, who assigned
not less than ten emission bands to Er-centred luminescence
upon ligand excitation in the [Er(quinolinate)3] complexes,9 a
behaviour that is reminiscent to that found for Er-doped
solids,12 we are not aware of the description of any erbium
coordination complex displaying a luminescence signal other
than the 1.5 μm band attributed to the Er(4I13/2→4I15/2) tran-
sition.13 This observation agrees with the large weighted
average effective vibrational energy of ∼2000 cm−1 participat-
ing in the non-radiative relaxation processes for lanthanide
complexes containing organic ligands.7 For ErIII, the reduced
energy gap p (in phonons units) between adjacent levels hardly
Scheme 4 Chemical structures of ligands L0–L2 and preparation of
the dinuclear [MLn(L1)3](CF3SO3)6 and trinuclear [MLnM(L1)3](CF3SO3)9
exceeds p = 2 and the dominant nonradiative relaxation pre- helicates (M = Ga, Cr and Ln = Ho, Er, Tm, Y).
This journal is © The Royal Society of Chemistry 2014
Dalton Trans.

View Article Online
Paper
Dalton Transactions
helicates could be isolated in 40–90% yields (Scheme 4 and
Table S1 in the ESI‡).
ESI-MS data unambiguously establish the charge
−
balance between [GaLn(L1)3]6+/6CF3SO3 and [GaLnGa(L2)3]9+/
−
9CF3SO3 (Tables S2–S3 in the ESI‡) whereas the 1H NMR
spectra collected for the diamagnetic Ga3+/Y3+ pair in aceto-
nitrile at submillimolar concentrations confirmed the self-
assembly of stable C3-symmetrical [GaY(L1)3]6+ (Fig. S1 in the
ESI‡) and D3-symmetrical [GaYGa(L2)3]9+ complexes (Fig. S2 in
the ESI‡), for which the unusual upfield chemical shifts of the
aromatic protons H6 and H9 together with the diastereotopic
splitting of H7 and H8 were diagnostic for the twisted arrange-
ment of the ligand strands in these triple-stranded helicates
(Fig. 1).15 Upon addition of one equivalent of paramagnetic
Eu(CF3SO3)3 into an acetonitrile solution of [GaYGa(L2)3]9+,
the 1H NMR spectrum remained unchanged for months at
room temperature, thus demonstrating the exceptional kinetic
inertness of these trinuclear helicates (Fig. S3 in the ESI‡). In
contrast, the trivalent lanthanide coordinated in the dinuclear
[GaY(L1)3]6+ complex was more accessible to metal exchange,
and the replacement of Y3+ with Eu3+ according to the equili-
Fig. 2 Kinetic traces monitored by 1H NMR for the reaction of [GaY
(L1)3](CF3SO3)6 (1.0 eq.) with Eu(CF3SO3)3 (1.0 eq.) at 293 K. The dotted
lines correspond to the best exponential fits.
brium (1) could be monitored on the timescale of hours (Fig. 2
and S4 in the ESI‡).
6þ
6þ
Y;Eu
exch
½GaYðL1Þ3ꢀ þ Eu3þ Ð ½GaEuðL1Þ3ꢀ þ Y3þ
K
ð1Þ
ð2Þ
Ga;Eu;L1
β
½GaEuꢀeq:½Yꢀeq:
1;1;3
Y;Eu
exch
K
¼
¼
Ga;Y;L1
½GaYꢀeq:½Euꢀeq:
β
1;1;3
The integration of the 1H NMR signals at chemical
Y,Eu
exch
equilibrium gave
K
= 134(3) according to eqn (2),
from
which
a
stabilization
factor
¼ 2:13ð1Þ
of
was
ꢀ
ꢁ
ꢀ
ꢁ
ꢀ
ꢁ
Y;Eu
Ga;Eu;L1
1;1;3
Ga;Y;L1
ξ ¼ log K
¼ log β
ꢁ log β
1;1;3
exch
deduced for the replacement of a small Y3+ with a midrange size
Eu3+ cation in the nine-coordinate cavity found in [GaLn(L1)3]6+.
Ga;Ln;L1
The β
stands for the cumulative formation constants of
1;1;3
the [GaLn(L1)3]6+ complexes, while [GaLn]eq. and [Ln]eq. corre-
spond to the equilibrium concentrations of complexes and
free metals (a standard concentration cθ = 1.0 m is taken for
the reference state, see Appendix 2 in the ESI‡). This trend is
Y→Eu
CN=9
remarkable when one considers the minute ΔR
= 0.045 Å
increase in the lanthanide nine-coordinate ionic radius, but it
is in line with ξ = 1.3(2) reported for the replacement of Ln =
Ho→Gd
CN=9
Ho3+ with Ln = Gd3+ in [Ln(L0)3]3+(ΔR
= 0.035 Å).14c
Since no significant quantity of intermediate complexes could
be detected by 1H NMR during the exchange reaction
(Fig. S4‡), reaction (1) could be tentatively considered as a
reversible second-order elementary reaction, for which the
quadratic rate law is given by
d½GaYꢀ
2
ꢁ
¼ ½GaYꢀ ðkf ꢁ krÞ þ ½GaYꢀfkfð½Euꢀtot ꢁ ½Gaꢀtot
þ krð½Yꢀtot þ ½GaꢀtotÞg ꢁ kr½Gaꢀtot½Yꢀtot
Þ
dt
ð3Þ
where kf and kr stand for the forward and backward second-
order rate constants, respectively (see Appendix 3 in the
ESI‡).16 However, the observation of mono-exponential kinetic
traces for both [GaY] depletion and [GaEu] rising with
a common rate constant k = 2.3(1) × 10−5 s−1 (τ1/2 = ln(2)/k =
8.3(4) hours, Fig. 2) suggests the more complicated mechan-
Fig. 1 1H NMR spectra with assignment for (a) L1 and [GaY(L1)3]6+ and
(b) L2 and [GaYGa(L2)3]9+ at 293 K (in CDCl3 for the ligands and in
CD3CN for the complexes).
Dalton Trans.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
ism shown in eqn (4), in which the first-order dissociation pro-
cesses characterized by k1 and k−2 values are the rate limiting
steps.17
k1
6þ
3þ
½GaYðL1Þ3ꢀ þ Eu3þ Ð ½GaðL1Þ3ꢀ þ Y3þ þ Eu3þ
k
ꢁ1
k2
6þ
Ð ½GaEuðL1Þ3ꢀ þ Y3þ
ð4Þ
k
ꢁ2
We conclude that the non-covalent cryptate [GaYGa(L2)3]9+
is sufficiently inert (τ1/2 > 1 month) to be used as a host matrix
for diluting isomorphous and photophysically active [GaErGa-
(L2)3]9+, [CrYCr(L2)3]9+ or [CrErCr(L2)3]9+ complexes using
co-crystallization. In contrast, [GaY(L1)3]6+ is too labile (τ1/2
=
8.3(4) hours) for accommodating [CrEr(L1)3]6+ guests without
undergoing lanthanide exchange processes.
Slow diffusion of tert-butylmethylether into an acetonitrile
solution of [GaErGa(L2)3]9+ provided prisms of [GaErGa(L2)3]-
(CF3SO3)9(CH3CN)35.5, the crystal structure of which (Fig. S5‡)
is isostructural with those reported for [CrLnCr(L2)3]
(CF3SO3)9(C2H5CN)30 (Ln = Eu, Yb, Fig. 3 and Table S4‡).10 The
close packing of the triple-helical cations [GaErGa(L2)3]9+
in the crystal produces large interstitial cavities, which are
filled with disordered counter-anions and solvent molecules
(Fig. S6–S7‡). The erbium atom is coordinated by nine hetero-
cyclic nitrogen atoms occupying the vertices of a distorted tri-
capped trigonal prism. We were unable to obtain X-ray quality
crystals for [GaEr(L1)3](CF3SO3)6, but the slow diffusion of tert-
butylmethylether into a concentrated propionitrile solution
of [CrEr(L1)3]6+ succeeded in crystallizing the analogous
dinuclear [CrEr(L1)3](CF3SO3)6(C3H5N)26 complex. The [CrEr-
(L1)3]6+ cations are packed in columns with opposite helicities
along the crystallographic threefold axes passing through
the metal ions (Fig. 4a, S8 and Table S5‡). In line with the
rigidity of the polyaromatic strands, the intermetallic distance
Cr⋯Er = 8.704(1) Å found in [CrEr(L1)3]6+ compares well with
those found in the trinuclear analogue (Ga1⋯Er1 = 8.974(1) Å
and Er1⋯Ga2 = 8.887(1)Å). The superimposition of the
triple-helical cations [CrEr(L1)3]6+ with [GaErGa(L2)3]9+
further demonstrates similar triple-helical organizations
(Fig. 4b).
Fig. 4 (a) ORTEP view with the numbering scheme of the asymmetric
unit in the crystal structure of [CrEr(L1)3](CF3SO3)6(C3H5N)26. Ellipsoids
are represented at the 30% probability level and hydrogen atoms are
omitted for clarity. (b) Superimposition of the molecular structures of
[CrEr(L1)3]6+ (green) and [GaErGa(L2)3]9+ (red).
Table 1 Characteristic excited state lifetimes (τ) and associated rate
constants (k = τ−1) measured for [CrLn(L1)3](CF3SO3)6 and [CrLnCr(L2)3]-
(CF3SO3)9 in the solid state at 10 Ka
[CrLn(L1)3]6+
[CrLnCr(L2)3]9+
τ1C→r 0(CrnY)/s
2.767(10) × 10−3
362(1)
2.336(10) × 10−3
428(2)
k1C→r 0(CrnY)/s−1
τ1E→r 0(GanEr)/s
4.28(2) × 10−6
2.34(1) × 105
3.8(4) × 10−8
2.6(3) × 107
1.615 × 103b
2.6(3) × 107c
295(5)
1.5(1) × 10−6
6.7(4) × 105
4.0(2) × 10−8
2.5(1) × 107
1.615 × 103b
2.5(1) × 107c
170(4)
kE1→r 0(GanEr)/s−1
τ2E→r 0(GanEr)/s
koEbrs(GanEr)/s−1
kE2→r 0(GanEr)/s−1
kE2→r 1(GanEr)/s−1
WC1r→Er(CrnEr)/s−1
a k1C→r 0, k1E→r 0 and koEbrs(GanEr) = 1/τ2E→r 0(GanEr) = kE2→r 1 + k2E→r 0 are the decay
rate constants (i.e. the sum of radiative and non-radiative processes) of
the Cr(2E→4A2) transitions measured in CrnY, and of the Er(4I13/2→4I15/2
)
and Er(4S3/2→4I15/2
) transitions measured in GanEr, respectively.
The photophysical properties confirmed the structural simi-
WC1r→Er are the rate constants for the Cr(2E)→Er(4I9/2) energy transfer
ꢂ
ꢃꢁ1
larities, and ligand-centred excitation at λex = 405 nm (ν
=
¯
exc
4S3=2
¼ ð619 μsÞꢁ1 is taken as the Er(4S3/2
)
b
2!0
processes.
k
¼
τEr;rad
Er
radiative lifetime measured for Er3+-doped into yttrium aluminium
garnet.21 c k2E→r 1(GanEr) = koEbrs(GanEr) − k
.
2→0
Er
24 691 cm−1) of solid-state samples of [GaEr(L1)3](CF3SO3)6
and [GaErGa(L2)3](CF3SO3)9 displayed a standard ligand-to-
erbium energy transfer followed by NIR emission signals
centred at 6500 cm−1 assigned to the split Er(4I13/2→4I15/2
)
transition.11 The roughly temperature-independent mono-
exponential lifetimes found for the Er(4I13/2) excited state
(3.1(4) μs for GaEr and 4.1(3) μs for GaErGa between 10 and
293 K, entries 3 and 4 in Table 1 and Fig. S9‡) combined with
the detection of additional green Er(4S3/2→4I15/2) emission
Fig. 3 Superimposition of the molecular structures of [GaErGa(L2)3]9+
(red) and [CrEuCr(L2)3]9+ (blue) observed in the crystal structures of
[GaErGa(L2)3](CF3SO3)9(CH3CN)35.5 and [CrEuCr(L2)3](CF3SO3)9(C2H5-
CN)30.
10 The hydrogen atoms are omitted for clarity.
This journal is © The Royal Society of Chemistry 2014
Dalton Trans.

View Article Online
Paper
Dalton Transactions
at 3 K in the solid state (entries 5 and 6 in Table 1, Fig. 5b and
S11‡).
In conclusion, the tight wrapping of the three bulky poly-
aromatic N-heterocyclic tridentate binding units in the triple-
helical [GaLn(L1)3](CF3SO3)6 and [GaLnGa(L2)3](CF3SO3)9 com-
plexes produces compact, stable and inert nine-coordinate cav-
ities for the lanthanide cations. For Ln = Er, the concomitant
existence of a comparatively long-lived intermediate Er(4I13/2
)
excited state and of a short-lived, but emissive Er(4S3/2) excited
state at higher-energy fulfils the two conditions required
for the implementation of ETU in Er-containing molecular
systems (Scheme 3). In a control experiment, direct continu-
ous-wave laser excitation of the Er(4I9/2←4I15/2) transition
centred at λexc = 647 nm (ν = 15 454 cm−1) failed to induce
¯
exc
upconverted signals in the Ga/Er complexes (Fig. S12‡), which
(i) confirmed Güdel’s conclusion that there was little chance of
detecting a single ion ESA in these molecular complexes,7 and
(ii) demonstrated that ligand-centred non-linear two-photon
absorption processes did not compete efficiently with ETU
under these conditions.
Optimizing the chromium coordination sphere for its use as a
sensitizer in ETU processes with an erbium activator
Based on the specific photophysical properties of the ErIII acti-
vator implemented in [MEr(L1)3](CF3SO3)6 and [MErM(L2)3]-
(CF3SO3)9, the associated M sensitizers should fit the following
criteria: They should (1) replace GaIII in order to yield a struc-
turally-similar, kinetically inert and thermodynamically stable
pseudo-octahedral [MN6] coordination building block, (2)
possess a sensitizing excited state, the energy of which is in
close resonance with one of the Er-centred excited state for
maximizing M→Er energy transfer processes and (3) display a
transparent optical window centred around 18 450 cm−1 in
order to avoid the quenching by non-radiative energy transfer
processes of the target upconverted emission arising from the
Er(4S3/2) excited state.11 Altogether, the trivalent chromium
with its [Ar]3d3 electronic configuration is a promising candi-
date since the energies of its six lower excited levels can
be tuned by the ligand-field strength (measured by Δ in
pseudo-octahedral complexes) and by the nephelauxetic effect
(measured by the reduction of the Racah parameters B and C
for electron–electron repulsion) produced by the terminal
didentate benzimidazole-2-yl-pyridine binding units involved
in the [CrN6] chromophores in [CrLn(L1)3](CF3SO3)6 and
[CrLnCr(L2)3](CF3SO3)9 (eqn (S1)–(S6) in Fig. S13‡).6 However,
CrIII is too inert to be used in a self-assembly process, and we
therefore resorted to more labile CrII precursors followed by
outer-sphere dioxygen oxidation for synthesizing [CrLn(L1)3]-
(CF3SO3)6 and [CrLnCr(L2)3](CF3SO3)9 (Ln = Y, Er) complexes
in 80–90% yields (Scheme 3 and Tables S1–S3‡). Orange X-ray
quality prisms of [CrYCr(L2)3](CF3SO3)9 were found to be iso-
structural with [GaErGa(L2)3](CF3SO3)9, [GaYGa(L2)3](CF3SO3)9,
[CrEuCr(L2)3](CF3SO3)9 and [CrYbCr(L2)3](CF3SO3)9.11 The
detailed analyses of the absorption and emission spectra
recorded for [CrY(L1)3](CF3SO3)6 and [CrYCr(L2)3](CF3SO3)9
with the help of eqn (S1)–(S6) gave Δ = 20 130 cm−1, B =
Fig. 5 (a) Luminescence spectra recorded for [GaEr(L1)3](CF3SO3)6 (the
black trace) and [GaErGa(L2)3](CF3SO3)9 (the blue trace) at 3 K in the
solid state (λex = 405 nm or νexc = 24 691 cm−1, cutoff filter 420 nm). The
¯
dips correspond to internal Er-centred re-absorption of ligand-centred
emission.13g (b) Time-resolved luminescence spectra recorded between
t = 0 and t = 350 ns (Δt = 10 ns) and luminescence decay trace (inset)
recorded for the Er(4S3/2→4I15/2) transition in [GaErGa(L2)3](CF3SO3)9 at
3 K in the solid state (λex = 405 nm or νexc = 24 691 cm−1, cut-off filter
¯
420 nm). The black curve corresponds to the best exponential fit
applied to the decay points deduced from the integration of the inten-
sity traces of the Er(4S3/2→4I15/2) transition.
centred at λem = 542 nm in both complexes (νem = 18 450 cm−1
,
¯
Fig. 5a and S10‡) reflect the exceptional protection of the
erbium cation from interactions with high-energy phonons
and vibrations in these [ErN9] chromophores.11,18 Though
technically challenging (see the Experimental section), the
kinetic traces of the Er-centred green luminescence could be
separated from the broad ligand-centred π*→π emission and
provided mono-exponential decay values with characteristic
Er(4S3/2) lifetimes of 38(4) ns for GaEr and 40(2) ns for GaErGa
Dalton Trans.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
Combining chromium sensitizers with erbium activators for
the implementation of ETU processes in molecular complexes
In the triple-stranded dinuclear [CrEr(L1)3](CF3SO3)6 and tri-
nuclear [CrErCr(L2)3](CF3SO3)9 complexes, the optimized nine-
coordinated pseudo-tricapped [ErN9] chromophore is located
at ca. 9 Å from one (dinuclear CrEr) or two (trinuclear CrErCr)
strong-field [CrN6] sensitizers (see Fig. 5b). This intermetallic
distance was designed for approaching the critical radius for
50% efficiency in the Cr→Er energy transfer,10 a situation
which maximizes the Cr-centred ETU mechanism operating in
the trinuclear CrErCr complex (Fig. 7b, green trace).11 Inter-
metallic communication via Cr(2E)→Er(4I9/2) energy transfer is
indeed evidenced by the 20–50% reduction of the Cr(2E)
excited lifetimes observed upon replacement of photophysi-
cally inactive Ln = Y (entries 1 and 2 in Table 1) with Ln = Er in
CrLn and CrLnCr complexes (Fig. S14‡).11 Assuming that the
Cr→Er energy transfer is the only source of additional quench-
ing of the Cr(2E) excited state going from Ln = Y to Ln = Er,
eqn (5) estimates the rate constants for Cr(2E)→Er(4I9/2) energy
transfers, which amount to W1
= 295(5) s−1 for CrEr and
Cr→Er
170(4) s−1 for CrErCr at 10 K (entry 9 in Table 1 and Fig. S15,‡
τCr stand for the characteristic lifetimes of the Cr(2E) levels
measured in the various complexes).11
Fig. 6 Jablonski diagrams for the [CrN6] chromophores in [CrY(L1)3]-
(CF3SO3)6 (left-hand side) and in [CrYCr(L2)3](CF3SO3)9 (right-hand side)
and for the [ErN9] chromophores in [GaEr(L1)3](CF3SO3)6 and [GaErGa-
(L2)3](CF3SO3)9 (centre). The energies of the excited levels deduced
from the absorption, excitation and emission spectra are given with
normal fonts, while the values obtained by computing are given in
italics. The emissive levels are shown with full downward arrows men-
tioning their characteristic lifetimes in pure solids at 3–10 K.11
obs
W1Cr!Er ¼ k ðCrnErÞ ꢁ kC1!r 0ðCrnYÞ
Cr
¼ ½τCrðCrnErÞꢀꢁ1 ꢁ ½τCrðCrnYÞꢀꢁ1
ð5Þ
The resulting ETU mechanism proposed in Scheme 3 can
be thus modelled with two simplified kinetic schemes
(Fig. 7).20 As suggested by these diagrams, NIR irradiation
into the Cr(2E,2T1←4A2) transitions at 700 ≤ λexc ≤ 750 nm
665 cm−1 and C = 2876 cm−1 for Cr3+ coordinated in the dinuc-
lear complex, and Δ = 19 940 cm−1, B = 736 cm−1 and C =
2737 cm−1 for Cr3+ in the trinuclear helicate (Fig. S13‡). These
parameters are comparable to those found for pseudo-octa-
hedral strong-field [Cr(2,2′-bipyridine)3]3+ (Δ = 23 240 cm−1, B =
761 cm−1 and C = 3044 cm−1),6 which indicates close locations
of the didentate benzimidazole-pyridine and 2,2′-bipyridine
binding units in the spectrochemical and nephelauxetic
series.
(13 330 ≤ ν ≤ 14 280 cm−1) indeed produced weak upcon-
¯
exc
verted green emissions arising from the Er(4S3/2→4I15/2) tran-
sitions (Fig. S16a‡), while the experimental slopes of 1.7–2.0
reported for the log–log plot of the upconverted intensity with
respect to the incident laser intensities confirmed the succes-
sive absorption of two photons (Fig. S16b‡).11
The complete set of kinetic rate constants collected in
Table 1 allows us to further explore the ETU mechanisms oper-
ating in the dinuclear (Fig. 7a) and trinuclear (Fig. 7b) com-
plexes. The macroscopic intensity measured for the final
upconverted signal in CrEr and CrErCr is proportional to the
normalized population density in the Er(4S3/2) excited state
The complete energy level diagram built for the [CrN6] and
[ErN9] chromophores in Fig. 6 shows (i) a good match between
the energy levels of the chromium sensitizer and those of the
Er 4S
l
erbium activator for implementing efficient Cr(2E)→Er(4I9/2
)
j ð 3=2Þ
N
modulated by its intrinsic emission quantum yield
. The ratio of the upconverted intensities recorded
Er 4S
l
j ð 3=2Þ
energy transfer processes in the target [CrEr(L1)3](CF3SO3)6
and [CrErCr(L2)3](CF3SO3)9 helicates, (ii) large Cr(4T2)–Cr(2T1)
energy gaps compatible with the unquenched emission arising
from the final Er(4S3/2) level and (iii) long luminescence life-
times for the intermediate Cr(2E) and Er(4I13/2) levels involved
in the planned ETU process (Scheme 3). As expected, continu-
ous-wave laser excitation of the Cr(2E,2T1←4A2) transitions in
the control compounds [CrY(L1)3](CF3SO3)6 and [CrYCr(L2)3]-
ϕlum
for the two complexes under similar experimental conditions
IuEpr (CrErCr)/IuEpr (CrEr) thus obeys eqn (6).11
Er 4S
l
Er 4S
l
j ð 3=2Þ
up
Er
j ð 3=2Þ
ϕlum
ðCrErCrÞ
ðCrErÞ
I
ðCrErCrÞ
N
ðCrErCrÞ
ðCrErÞ
¼
ꢂ
ð6Þ
up
Er 4S
l
Er 4S
l
j ð 3=2Þ
j ð 3=2Þ
I
ðCrErÞ
Er
N
ϕlum
Since the Er3+ cation is only slightly more accessible to
(CF3SO3)9 (700 ≤ λexc ≤ 750 nm, 13 330 ≤ ν ≤ 14 280 cm−1
)
¯
exc
high-energy vibrations in [CrEr(L1)3]6+ than in [CrErCr(L2)3]9+
failed to induce upconverted signals (Fig. S16a‡), which
excludes the possibilities of Cr-centred upconversion processes
or spurious ETU from impurities.19
(vide
supra),
one
can
reasonably
assume
that
Er 4S
l
Er 4S
l
j ð 3=2Þ
j ð 3=2Þ
ϕlum
ðCrErCrÞ ꢃ ϕlum
ðCrErÞ and that therefore the
This journal is © The Royal Society of Chemistry 2014
Dalton Trans.

View Article Online
Paper
Dalton Transactions
section of the Cr(2E,2T1←4A2) transition, h is the Planck’s con-
stant and c is the vacuum speed of light.22
λp
excð0!1Þ
0!1
k
¼
Pσ
ð7Þ
Cr
Cr
hc
Assuming that all the kinetic rate constants involved in the
kinetic diagrams depicted in Fig. 7 are in hand, the normal-
ized population densities for each level can be computed with
eqn (8), whereby M is the kinetic matrix associated with each
kinetic diagram (Fig. S17–S18‡).11,17
ꢄ
ꢅ
h
i
dNjil
dt
¼ M ꢄ Njil
ð8Þ
The steady-state regime dN|i>/dt = 0 induced by the continu-
ous-wave irradiation of the sensitizer reduces eqn (8) to M ×
[N|i>] = 0, a situation ‘easily’ solved for [N|i>] by using eqn (9),
where M′ is the rectangular matrix produced by the inclusion
P
of mass conservation Njil ¼ Ntot ¼ 1 into the square kinetic
matrix M (TM′ is the transpose matrix, see Appendix 4‡ for
i
mathematical details).11
2
3
0
ꢂ
h
i
6
6
4
7
7
5
ꢁ1
T
T
Njil ¼ ð M′ ꢄ M′Þ ꢄ M′ ꢄ
ð9Þ
0
Ntot
The successive introduction of the rate constants
gathered for [CrEr(L1)3](CF3SO3)6 and for [CrErCr(L2)3]
(CF3SO3)9 into eqn (9) provides the normalized population
Er 4S
l
Er 4S
l
j ð 3=2Þ
j ð 3=2Þ
densities and their ratio N
ðCrErCrÞ=N
ðCrErÞ
so that the two experimentally non-accessible parameters, i.e.
the Cr(2E,2T1←4A2) absorption cross sections σ
required
0→1
Cr
exc(0→1)
Cr
for computing k
in eqn (7) and the rate constants
for the second energy transfer process WC2r→Er responsible
for the Er(4S3/2←4I13/2
superexcitation pathway, can be
)
estimated. Assuming that (1) σ0C→r 1(CrEr) ≈ σC0→r 1(CrErCr) =
10−24 m2,23 and (2) the trend in energy transfer
processes experimentally observed for WC1r→Er also holds
for WC2r→Er (i.e. W2Cr→Er(CrEr) = W1Cr→Er(CrEr) = 295 s−1 and
WC2r→Er(CrErCr) = WC1r→Er(CrErCr) = 170 s−1), we calculated
Fig. 7 Simplified kinetic schemes20 depicting the modelling of energy
transfer upconversion (ETU) processes occurring upon off-resonance
irradiation into the sensitizer-centred absorption bands in (a) the dinuc-
Er 4S
l
Er 4S
l
j ð 3=2Þ
j ð 3=2Þ
N
ðCrErCrÞ=N
ðCrErÞ ꢃ 10 (Fig. 8), in fair agree-
ment with the experimental ratios of IuEpr (CrErCr)/IEupr (CrEr) =
4–10 reported for the upconverted intensities measured for the
two complexes in the 10–50 K range (Fig. S16‡).11 For large
pump intensities (log(P) ≥ 1), the computed IuEpr (CrErCr)/
IuEpr (CrEr) ratio takes a downturn due to saturation effects
affecting the Cr-centred ETU mechanism in the triad (Fig. 8b).
lear [CrEr(L1)3]6+ and (b) the trinuclear [CrErCr(L2)3]9+ complexes.
exc(0→1)
Cr
k
is the pumping rate constant for irradiation into the Cr(2T1←4A2)
1→0
Cr
1→0
absorption band at λp = 705 nm (eqn (9)). k , respectively k
k2E→r 1 and k
,
Er
2→0
are the decay rate constants (i.e. the sum of radiative and
Er
non-radiative processes) for the Cr(2E→4A2) transition in CrnY and for
the Er(4I13/2→4I15/2), Er(4S3/2→4I13/2) and Er(4S3/2→4I15/2) transitions in
GanEr, respectively. WC1 r→Er and W2Cr→Er are the rate constants of Cr→Er
energy transfer processes. The red dashed pathways highlight the Er-
centred ETU mechanism, while the green dashed pathway holds for the
additional Cr-centred ETU mechanism operating in CrnEr when n ≥ 2.11
Conclusion
This work demonstrates that the photophysically active and
kinetically inert CrIII cations found in dinuclear [CrLn(L1)3]6+
and trinuclear [CrLnCr(L2)3]9+ triple-stranded helicates can be
replaced with the closed-shell GaIII to give slightly less inert,
isostructural [GaLn(L1)3]6+ and trinuclear [GaLnGa(L2)3]9+
complexes, for which the photophysical properties of the
ratio of intensities roughly corresponds to the ratio of the nor-
Er 4S
l
Er 4S
l
j ð 3=2Þ
j ð 3=2Þ
malized population densities N
ðCrErCrÞ=N
ðCrErÞ.
Under Cr-centred excitation at a fixed incident pump intensity
exc(0→1)
P, the pumping rate constant k
is given by eqn (7),
is the absorption cross
Cr
0→1
Cr
where λp is the pump wavelength, σ
Dalton Trans.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
centred NIR excitation induces light-upconversion obeying an
ETU mechanism (Scheme 3). Such behaviour requires a collec-
tion of favourable conditions which are difficult to gather in a
single entity and, despite efficient Cr(2E)→Tm(3H4) energy
transfer processes observed in [CrTmCr(L2)3]9+ (Fig. S22‡), we
were unable to detect chromium-to-thulium ETU (Fig. S23‡),
probably because the target blue Tm(1G4→3H6) emission is
quenched by an efficient Tm(1G4)→Cr(4T2) energy back-trans-
fer (Fig. S21‡). For [CrHoCr(L2)3]9+, the poor energy match
between the Cr(2E) and Ho(5I5) levels (Fig. S21‡) limits the
Cr→Ho energy transfer rate constant to less than 100 s−1
(Fig. S22‡), and no upconversion could be evidenced upon
Cr-centred excitation (Fig. S24‡). Finally, numerical simu-
lations with eqn (9) show that using reasonable values for the
various excited lifetimes in metallic sensitizers (in the milli-
second range) and activators (in the microsecond range) involved
in heterometallic MnLn (supra)molecular complexes, both Ln-
centred (the red pathway in Fig. 7) and M-centred (the green
pathway in Fig. 7) ETU mechanisms greatly benefit from the
maximization of (i) the number n of sensitizers per activator
and (ii) the rate constants for the second M→Ln energy trans-
fer process (WM2 →Ln). As a test, a unique and common value
for the intermetallic Cr→Er energy transfer rates constants in
the CrEr and CrErCr complexes was considered for a compari-
son purpose. Introducing WC2r→Er(CrEr) = WC1r→Er(CrEr) =
WC2r→Er(CrErCr) = WC1r→Er(CrErCr) = 170 s−1 into eqn (9) yields
Er 4S
l
Er 4S
l
j ð 3=2Þ
j ð 3=2Þ
N
ðCrErCrÞ=N
ðCrErÞ ꢃ 220 (Fig. S25‡), a ratio
which would have prevented the experimental detection of
upconverted luminescence for the dinuclear CrEr helicate with
our setup. The weak, but measurable upconverted signal
experimentally recorded for the latter [CrEr(L1)3]6+ complex
upon chromium-centred NIR excitation is thus indebted to
the Cr→Er rate constant, WCr→Er, which is twice as large as
that measured in the trinuclear [CrErCr(L2)3]9+ analogue.
Further optimization should therefore focus on the increase of
WM2 →Ln with the help of super-exchange mechanisms in poly-
nuclear MnLn complexes.24
Fig. 8 (a) Log–log plot and (b) ratio of the quadratic dependence of the
4
Er S3=2
l
j
ð
Þ
steady-state normalized population densities N
on the incident
pump intensity (P in W mm−2) computed for the dinuclear [CrEr(L1)3]-
(CF3SO3)6 (the full trace) and for the trinuclear [CrErCr(L2)3](CF3SO3)9
(the dotted trace) complexes using eqn (9) and the kinetic rate constants
gathered in Table 1. σ0C→r 1(CrEr) ≈ σ0C→r 1(CrErCr) = 10−24 m2,23 W2Cr→Er(CrEr)
= WC1 r→Er(CrEr) = 295 s−1 and WC2 r→Er(CrErCr) = W1Cr→Er(CrErCr) = 170 s−1
(see the text).
lanthanide cations can be deciphered without the compli-
cation due to the presence of the d-block cations. In this
context, the pseudo-tricapped trigonal prismatic [ErN9]
chromophore appears to be exceptional in its capacity to sim-
ultaneously emit downshifted visible light arising from the
Er(4S3/2) level and near-infrared emission from the lowest
Er(4I15/2) excited level upon ligand-centred excitation, and this
despite the presence of high-energy oscillators brought by the
organic ligand strands. It is therefore not so surprising, but
still remarkable, that the replacement of ErIII with either TmIII
(Fig. S19‡) or HoIII (Fig. S20‡) in [GaLnGa(L2)3]9+ similarly pro-
duces rich lanthanide-centred luminescence upon UV-exci-
tation of the aromatic receptor (Fig. S21‡). The combination of
the [ErN9] chromophore, acting as the activator, with optically-
adapted strong-field pseudo-octahedral [CrN6] sensitizers in
[CrEr(L1)3]6+ and [CrErCr(L2)3]9+ complexes further tolerates
radiative visible de-excitation of the Er(4S3/2) level, and Cr-
Experimental
Solvents and starting materials
These were purchased from Strem, Acros, Fluka AG and
Aldrich and used without further purification unless otherwise
stated. The trifluoromethanesulfonate salts Ln(CF3SO3)3·xH2O
were prepared from the corresponding oxide (Aldrich
99.99%).25 The synthesis of ligand L1 is detailed in Appendix
1.‡ The ligand L226 and the complexes [MLn(L1)3](CF3SO3)6·
x(H2O)·y(C3H5N) and [MLnM(L2)3](CF3SO3)9·x(H2O)·y(C3H5N)
were prepared according to literature procedures (M = Cr, Ga
and Ln = Ho, Tm, Er, Y using propionitrile as a solvent) and
characterized by elemental analyses (Table S1‡) and single-
crystal X-ray diffraction.11 Acetonitrile and dichloromethane
were distilled over calcium hydride.
This journal is © The Royal Society of Chemistry 2014
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Spectroscopic and analytical measurements
under vacuum and dissolved the resulting materials into
CD3CN. The 1H NMR spectra showed the presence of
the [GaErGa(L2)3]9+, resp. [CrEr(L1)3]6+ cations together with
acetonitrile and the resp. propionitrile molecules, but no
signal for the heavy ether. It was therefore assumed that the
crystals contained [GaErGa(L2)3](CF3SO3)9 and [CrEr(L1)3]-
(CF3SO3)6 together with nitrile solvent molecules. Summaries
of crystal data, intensity measurements and structure refine-
ments for [GaErGa(L2)3](CF3SO3)9(CH3CN)35.5 (16) and [CrEr-
(L1)3](CF3SO3)6(C3H5N)26 (17) are collected in Tables S4 and S9
(ESI‡). Each crystal was mounted on a kapton loop with protec-
tion oil. Cell dimensions and intensities were measured at
150 K on an Agilent Supernova diffractometer with mirror-
monochromated Cu[Kα] radiation (λ = 1.54 187) for 16 or
at 100 K using the Swiss-Norwegian beamlines, European
Synchtron Radiation Facility (λ = 0.8231 Å) for 17. Data were
corrected for Lorentz and polarization effects and for absorp-
tion. The structure was solved by direct methods (SIR97),27
and all other calculations were performed with ShelX9728
systems and ORTEP29 programs. CCDC-1003567 and
CCDC-1003568 contain the supplementary crystallographic
data.
The details of the setup and procedures used for recording
high-resolution emission spectra at variable temperatures
under various excitation conditions are reported in ref. 11.
Time-resolved luminescence spectra of [GaEr(L1)3](CF3SO3)6
and [GaErGa(L2)3](CF3SO3)9 were obtained from powder
samples directly mounted onto copper plates using conductive
silver glue, and cooled in an optical closed-cycle cryostat
capable of reaching low temperatures down to 3 K under a
helium atmosphere (Sumitomo SHI-950/Janis Research
CCS-500/204). High-resolution emission spectra were recorded
upon excitation with Nd:YAG lasers (Quantel Brilliant B) by
using the third harmonic at 355 nm. The laser was connected
to a gated CCD detector (Andor iSTAR) during the monitoring
of the luminescence spectra. The emitted light was analysed at
90° using an Andor SR-163 monochromator (slits 100 μm) with
a holographic grating (600 grooves mm−1, blazed at 500 nm).
Appropriate filters (395 or 435 nm cutoff) were used to remove
the residual excitation laser light, the Rayleigh scattered light
and associated harmonics from the emission spectra. A kinetic
series was recorded using a gate pulse width of 15 ns and by
scanning the delay from just before the arrival of the laser
pulse to 350 ns in steps of 10 ns. Each spectrum corresponds
to the sum of 1200–5000 scans. All the luminescence spectra
were transferred to a PC for data analysis. The intensity at
541 nm and the integrated intensity between 535 and 560 nm
corresponding to the Er(4S3/2→4I15/2) transition were calculated
for each curve after background correction of the ligand-
centred emission. The maximum intensity and the integrated
intensity vs. the time (0–350 ns) were plotted, from which the
characteristic lifetimes for the Er(4S3/2) level could be esti-
Warning. The quality of the diffraction data and their
refinement is low due to considerable disorder. For 16, twelve
triflate anions could be located and refined, but the absence of
features that could be unambiguously attributed to the
missing six triflate anions and to the solvent molecules in the
Fourier difference map forced us to apply the Squeeze/bypass
method (program PLATON) in order to take care of the remain-
ing electronic density. There is a huge void in the structure, in
which 2003 electrons per formula unit were attributed to
Squeeze. This corresponded to six triflates together with 71
acetonitrile molecules which were included in the unit cell. In
1
mated. H and 13C NMR spectra were recorded at 293 K on a
BrukerAvance 400 MHz spectrometer. Chemical shifts
are given in ppm with respect to tetramethylsilane Si(CH3)4.
Pneumatically-assisted electrospray (ESI) mass spectra
were recorded from 10−4 M solutions on an Applied Bio-
systems API 150EX LC/MS System equipped with a Turbo
Ionspray source®. Elemental analyses were performed by
K. L. Buchwalder from the Microchemical Laboratory of the
University of Geneva. Electronic absorption spectra in the
UV-Vis were recorded at 20 °C from solutions with a Perkin-
Elmer Lambda 900 spectrometer using quartz cells of 10 or
−
17, the six CF3SO3 counter-anions and interstitial solvent
molecules were highly disordered in large voids in the struc-
ture. In the absence of diagnostic features in the Fourier differ-
ence map, the Squeeze/bypass method (program PLATON)30
was used to take care of the remaining electron density, which
was eventually assigned to six triflates and 26 propionitrile
molecules (details are given in the associated CIF file).
1 mm path length. The mathematical analyses were performed Acknowledgements
by using Igor Pro® (WaveMetrics Inc.) and Excel® (Microsoft)
softwares.
This work was supported through grants from the Swiss
National Science Foundation (grant numbers 200020_140222
and 200020-125175), the University of Geneva (INNOGAP), la
Ligue contre le Cancer and the Institut National de la Santé et
X-Ray crystallography
Fragile crystals of [GaErGa(L2)3](CF3SO3)9(CH3CN)35.5 (16) de la Recherche Médicale (INSERM). The work was carried out
were obtained by slow diffusion of tert-butylmethylether into within the COST Actions TD1004 and CM1006.
an acetonitrile solution of [GaErGa(L2)3]9+, while those of
[CrEr(L1)3](CF3SO3)6(C3H5N)26 (17) were obtained by the slow
diffusion of tert-butylmethylether into a concentrated propio-
Notes and references
nitrile solution of [CrEr(L1)3]6+. As the crystals did not survive
when separated from their mother liquors, we deposited the
crystals on a filter paper, briefly sucked the excess solvent
1 (a) Nonlinear Optical Properties of Organic Molecules and
Crystals, ed. D. S. Chemla and J. Zyss, Academic Press,
Dalton Trans.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
Orlando, 1987; (b) P. N. Prasad and D. J. Williams, Introduc-
tion to Nonlinear Optical Effects in Molecules and Polymers,
John Wiley & Sons, Inc., New York, 1991; (c) J. Zyss, Mole-
Phys. B, 2006, 82, 289–294; (c) F. Prudenzano, L. Mescia,
L. Allegretti, V. Moizan, V. Nazabal and F. Smektala, Opt.
Mater., 2010, 33, 241–245.
cular Nonlinear Optics: Materials, Physics and Devices, Aca- 13 (a) J. W. Hofstraat, M. P. Oude Wolbers, F. C. J. M. Van
demic Press, Boston, 1994; (d) Nonlinear Optics of Organic
Molecules and Polymers, ed. H.S. Nalwa and S. Miyata, CRC
Press, Boca Raton, 1997.
Veggel, D. N. Reinhoudt, M. H. V. Werts and
J. W. Verhoeven, J. Fluoresc., 1998, 8, 301–308;
(b) M. P. Oude Wolbers, F. C. J. M. van Veggel,
F. G. A. Peters, E. S. E. van Beelen, J. W. Hofstraat,
F. A. J. Geurts and D. N. Reinhoudt, Chem. – Eur. J., 1998,
4, 772–780; (c) L. H. Sloof, A. Polman, M. P. O. Wolbers,
F. C. J. M. van Veggel, D. N. Reinhoudt and J. W. Hofstraat,
J. Appl. Phys., 1998, 83, 497–503; (d) B. M. Flanagan,
P. V. Bernhardt, E. R. Krausz, S. R. Lüthi and M. J. Riley,
Inorg. Chem., 2001, 40, 5401–5407; (e) K. Kuriki, Y. Koike
and Y. Okamoto, Chem. Rev., 2002, 102, 2347–2356;
(f) L. H. Sloof, A. van Blaaderen, A. Polman, G. A. Hebbink,
S. I. Klink, F. C. J. M. Van Veggel, D. N. Reinhoudt and
J. W. Hofstraat, J. Appl. Phys., 2002, 91, 3955–3980;
(g) V. S. Sastri, J.-C. Bünzli, V. R. Rao, G. V. S. Rayudu and
J. R. Perumareddi, Modern Aspects of Rare Earths and Their
Complexes, Elsevier, Amsterdam, 2003, ch. 8, pp. 629–632;
(h) S. Quici, G. Marzanni, A. Forni, G. Accorsi and
F. Barigelletti, Inorg. Chem., 2004, 43, 1294–1301;
(i) G. M. Davies, H. Adams, S. J. A. Pope, S. Faulkner and
M. D. Ward, Photochem. Photobiol. Sci., 2005, 4, 829–834;
( j) G. Mancino, A. J. Ferguson, A. Beeby, N. J. Long and
T. S. Jones, J. Am. Chem. Soc., 2005, 127, 524–525;
(k) J. B. Oh, Y. H. Kim, M. K. Nah and H. K. Kim, J. Lumin.,
2005, 111, 255–264; (l) M.-K. Nah, H.-G. Cho, H.-J. Kwon,
Y.-J. Kim, C. Park, H. K. Kim and J.-G. Kang, J. Phys. Chem.
A, 2006, 110, 10371–10374; (m) S. Destri, M. Pasini,
W. Porzio, F. Rizzo, G. Dellepiane, M. Ottonelli, G. Musso,
F. Meinardi and L. Veltri, J. Lumin., 2007, 127, 601–610;
(n) X. Li, Z. Si, C. Pan, L. Zhou, Z. Li, X. Li, J. Tang and
H. Zhang, Inorg. Chem. Commun., 2009, 12, 675–677;
(o) F. Artizzu, M. L. Mercuri, A. Serpe and P. Deplano,
Coord. Chem. Rev., 2011, 255, 2514–2529; (p) D. J. Lewis,
F. Moretta, A. T. Holloway and Z. Pikramenou, Dalton
Trans., 2012, 41, 13138–13146; (q) P. Martin-Ramos,
C. Coya, A. L. Alvarez, M. Ramos Silva, C. Zaldo,
J. A. Paixao, P. Chamorro-Posada and J. Martin-Gil, J. Phys.
Chem. C, 2013, 117, 10020–10030.
2 (a) F. Terenziani, C. Katan, E. Badaeva, S. Tretiak and
M. Blanchard-Desce, Adv. Mater., 2008, 20, 4641–4678;
(b) C. Andraud and O. Maury, Eur. J. Inorg. Chem., 2009,
4357–4371; (c) P. Neveu, D. K. Sinha, P. Kettunen, S. Vriz,
L. Jullien and D. Bensimon, Springer Ser. Chem. Phys., 2010,
96, 305–316; (d) A. D’Aléo, A. Bourdolle, S. Brustlein,
T. Fauquier, A. Grichine, A. Duperray, P. L. Baldeck,
C. Andraud, S. Brasselet and O. Maury, Angew. Chem., Int.
Ed., 2012, 51, 6622–6625.
3 (a) S. V. Eliseeva and J.-C. G. Bünzli, New J. Chem., 2011, 35,
1165–1176; (b) X. Huang, S. Han, W. Huang and X. Liu,
Chem. Soc. Rev., 2013, 42, 173–201; (c) J.-C. G. Bünzli and
A.-S. Chauvin, in Handbook on the Physics and Chemistry
of Rare Earths, ed. J.-C. G. Bünzli and V. K. Pecharsky,
Elsevier, North Holland, Amsterdam, 2014, vol. 44, pp.
169–281.
4 (a) T. N. Singh-Rachford and F. N. Castellano, Coord. Chem.
Rev., 2010, 254, 2560–2573; (b) J. Zhao, S. Ji and H. Guo,
RSC Adv., 2011, 1, 937–950; (c) A. Monguzzi, R. Tubino,
S. Hoseinkhani, M. Campione and F. Meinardi, Phys.
Chem. Chem. Phys., 2012, 14, 4322–4332; (d) J.-H. Kim,
F. Deng, F. N. Castellano and J.-H. Kim, Chem. Mater.,
2012, 24, 2250–2252; (e) X. Huang, S. Han, W. Huang and
X. Liu, Chem. Soc. Rev., 2013, 42, 173–201.
5 (a) F. Auzel, Chem. Rev., 2004, 104, 139–173; (b) B. M. van
der Ende, L. Aarts and A. Meijerink, Phys. Chem. Chem.
Phys., 2009, 11, 11081–11095; (c) M. Haase and H. Schäfer,
Angew. Chem., Int. Ed., 2011, 50, 5808–5829.
6 L. Aboshyan-Sorgho, M. Cantuel, S. Petoud, A. Hauser and
C. Piguet, Coord. Chem. Rev., 2012, 256, 1644–1663.
7 C. Reinhard and H. U. Güdel, Inorg. Chem., 2002, 41, 1048–
1055.
8 O. A. Blackburn, M. Tropiano, T. J. Sorensen, J. Thom,
A. Beeby, L. M. Bushby, D. Parker, L. S. Natrajan and
S. Faulkner, Phys. Chem. Chem. Phys., 2012, 14, 13378–
13384.
14 (a) C. Piguet, J.-C. G. Bünzli, G. Bernardinelli and
A. F. Williams, Inorg. Chem., 1993, 32, 4139–4149;
(b) C. Piguet, J.-C. G. Bünzli, G. Bernardinelli, C. G. Bochet
and P. Froidevaux, J. Chem. Soc., Dalton Trans., 1995, 83–97;
(c) S. Petoud, J.-C. G. Bünzli, F. Renaud, C. Piguet,
K. J. Schenk and G. Hopfgartner, Inorg. Chem., 1997, 36,
5750–5760.
9 H. Suzuki, Y. Nishida and S. Hoshino, Mol. Cryst. Liq.
Cryst., 2003, 406, 27/[221]–237/[231].
10 L.
Aboshyan-Sorgho,
C.
Besnard,
P.
Pattison,
K. R. Kittilstved, A. Aebischer, J.-C. G. Bünzli, A. Hauser
and C. Piguet, Angew. Chem., Int. Ed., 2011, 50, 4108–
4112.
11 Y. Suffren, D. Zare, S. V. Eliseeva, L. Guénée, 15 (a) C. Piguet, E. Rivara-Minten, G. Hopfgartner and
H. Nozary, T. Lathion, L. Aboshyan-Sorgho, S. Petoud,
A. Hauser and C. Piguet, J. Phys. Chem. C, 2013, 117,
26957–26963.
J.-C. G. Bünzli, Helv. Chim. Acta, 1995, 78, 1541–1566;
(b) C. Piguet, J.-C. G. Bünzli, G. Bernardinelli,
G. Hopfgartner, S. Petoud and O. Schaad, J. Am. Chem. Soc.,
1996, 118, 6681–6697.
12 (a) G. A. Kumar, R. Riman, S. C. Chae, Y. N. Jang, I. K. Bae
and H. S. Moon, J. Appl. Phys., 2004, 95, 3243–3249; 16 D. Lavabre, V. Pimienta, G. Levy and J. C. Micheau, J. Phys.
(b) A. K. Singh, S. B. Rai, D. K. Rai and V. B. Singh, Appl.
Chem., 1993, 97, 5321–5326.
This journal is © The Royal Society of Chemistry 2014
Dalton Trans.

View Article Online
Paper
Dalton Transactions
17 M. E. Starzak, Mathematical Methods in Chemistry 23 The absorption cross section σi→j (in cm2) is related to the
and Physics, Plenum press, New York, 1989, pp. 289–
357.
18 Because of the intense visible luminescence of [GaEuGa-
(L2)3](CF3SO3)9, the Er salts used for preparing [GaErGa-
molar extinction coefficient εi→j (in m−1 cm−1) by using
σ
i→j = 3.8 × 10−21·εi→j. Typical values of σCr ≈ 10−24 m2 have
been reported in ref. 19 for the Cr(2T1,2E←4A2) transitions
observed for Cr3+ doped in Cs2NaScCl6.
(L2)3](CF3SO3)9 must have the highest possible purity. 24 M. E. von Arx, E. Burattini, A. Hauser, L. van Pieterson,
Er2O3 (99.99%) provided the emission spectrum depicted
in Fig. 5 (the blue trace), but the use of Er2O3 (99.9%)
R. Pellaux and S. Decurtins, J. Phys. Chem. A, 2000, 104,
883–893.
showed concomitant Er- and Eu-centred emissions at low 25 J. F. Desreux, in Lanthanide Probes in Life, Chemical and
temperature (Fig. S10b‡).
19 O. S. Wenger, H. Güdel and S. Kück, J. Chem. Phys., 2002,
117, 909–913.
Earth Sciences, ed. G. Choppin and J.-C. G. Bünzli, Elsevier,
Amsterdam, 1989, ch. 2.
26 C. Piguet, B. Bocquet and G. Hopfgartner, Helv. Chim. Acta,
1994, 77, 931–942.
20 For the sake of clarity, the kinetic diagrams collected in
Fig. 7 are limited to the lower excited states involved for 27 A. Altomare, M. C. Burla, M. Camalli, G. Cascarano,
two successive Cr-centred excitations. However, all calcu-
lations take into account the complete schemes described
C. Giacovazzo, A. Guagliardi, G. Moliterni, G. Polidori and
R. Spagna, J. Appl. Crystallogr., 1999, 32, 115–119.
in ref. 11, which are reproduced in Fig. S17a–S18a‡ for the 28 G. M. Sheldrick, SHELXL97 Program for the Solution and
convenience of the reader.
21 S. Georgescu and V. Lupei, J. Quantum Electron., 1998, 34,
1031–1040.
Refinement of Crystal Structures, University of Göttingen,
Germany, 1997.
29 ORTEP3 for Windows. L. J. Farrugia, J. Appl. Crystallogr.,
1997, 30, 565.
22 M. Pollnau, D. R. Gamelin, S. R. Lüthi and
H. U. Güdel, Phys. Rev. B: Condens. Matter, 2000, 61, 3337– 30 P. Van der Sluis and A. L. Spek, Acta Crystallogr., Sect. A:
3346.
Fundam. Crystallogr., 1990, 46, 194–201.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2014