Optical properties of trinuclear metal chalcogenolate complexes-room temperature NIR fluorescence in [Cu2Ti(SPh)6(PPh3)2]

Article abstract of DOI:10.1039/c6dt04287b

The optical properties of four isostructural trinuclear chalcogenolato bridged metal complexes [Cu₂Sn(SPh)₆(PPh₃)₂], [Cu₂Sn(SePh)₆(PPh₃)₂], [Ag₂Sn(SPh)₆(PPh₃)₂] and [Cu₂Ti(SPh)₆(PPh₃)₂] have been investigated by absorption and photoluminescence spectroscopy and time-dependent density functional theory (TDDFT) calculations. All copper-tin compounds demonstrate near-infrared (NIR) phosphorescence at ～900-1100 nm in the solid state at low temperature, which is nearly absent at ambient temperature. Stokes shifts of these emissions are found to be unusually large with values of about 1.5 eV. The copper-titanium complex [Cu₂Ti(SPh)₆(PPh₃)₂] also shows luminescence in the NIR at 1090 nm but with a much faster decay (τ ～ 10 ns at 150 K) and a much smaller Stokes shift (ca. 0.3 eV). Even at 295 K this fluorescence is found to comprise a quantum yield as high as 9.5%. The experimental electronic absorption spectra well correspond to the spectra simulated from the calculated singlet transitions. In line with the large Stokes shifts of the emission spectra the calculations reveal for the copper-tin complexes strong structural relaxation of the excited triplet states whereas those effects are found to be much smaller in the case of the copper-titanium complex.

Full text of DOI:10.1039/c6dt04287b

Dalton
Transactions
View Article Online
View Journal
PAPER
Optical properties of trinuclear metal
chalcogenolate complexes – room temperature
NIR fluorescence in [Cu2Ti(SPh)6(PPh3)2]†
Cite this: DOI: 10.1039/c6dt04287b
Michael Kühn,a Sergei Lebedkin,b Florian Weigend*b,a and Andreas Eichhöfer*b,c,d
The optical properties of four isostructural trinuclear chalcogenolato bridged metal complexes [Cu2Sn
(SPh)6(PPh3)2], [Cu2Sn(SePh)6(PPh3)2], [Ag2Sn(SPh)6(PPh3)2] and [Cu2Ti(SPh)6(PPh3)2] have been investi-
gated by absorption and photoluminescence spectroscopy and time-dependent density functional
theory (TDDFT) calculations. All copper–tin compounds demonstrate near-infrared (NIR) phosphor-
escence at ∼900–1100 nm in the solid state at low temperature, which is nearly absent at ambient temp-
erature. Stokes shifts of these emissions are found to be unusually large with values of about 1.5 eV. The
copper–titanium complex [Cu2Ti(SPh)6(PPh3)2] also shows luminescence in the NIR at 1090 nm but with
a much faster decay (τ ∼ 10 ns at 150 K) and a much smaller Stokes shift (ca. 0.3 eV). Even at 295 K this
fluorescence is found to comprise a quantum yield as high as 9.5%. The experimental electronic absorp-
tion spectra well correspond to the spectra simulated from the calculated singlet transitions. In line with
the large Stokes shifts of the emission spectra the calculations reveal for the copper–tin complexes
strong structural relaxation of the excited triplet states whereas those effects are found to be much
smaller in the case of the copper–titanium complex.
Received 10th November 2016,
Accepted 5th January 2017
DOI: 10.1039/c6dt04287b
www.rsc.org/dalton
emission efficiency (up to ∼90% in the solid state at ambient
temperature) are considered as promising candidates for
Introduction
Copper complexes form a class of coordination compounds future economical and mass-producible organic light-emitting
with very rich photochemical and photophysical behavior. In diodes (OLEDs)8,9 and light-emitting electrochemical cells
particular, numerous copper-based complexes have been (LECs).10,11 Additional diversity of photophysical properties
found to show bright photoluminescence (PL).1–7 Depending can be introduced by polynuclear structures. For instance,
on the specific complex structure, the PL (typically phosphor- we have recently demonstrated that the polynuclear copper
escence or delayed fluorescence) has been observed across the complexes [Cu12S6(dpppt)4] and [Cu12S6(dppo)4] attain PL
whole visible range from blue to deep-red and it can demon- efficiency at ambient temperature larger than 48%, which is
strate various electronic relaxation and decay patterns in the complemented by high photostability.12,13 Consequently, these
solid state and solution. Based on the abundant materials, compounds may be a valuable addition to the line of copper-
several types of mononuclear Cu(I) complexes with a high based luminophors. Furthermore, introduction of another
type of metal provides supplementary structural and photo-
physical variability. In this respect, it has been shown recently
aInstitut für Physikalische Chemie, Abteilung für Theoretische Chemie, Karlsruher
Institut für Technologie (KIT), Campus Süd, Fritz-Haber-Weg 2, 76131 Karlsruhe,
Germany
bInstitut für Nanotechnologie, Karlsruher Institut für Technologie (KIT), Campus
Nord, Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-Leopoldshafen, Germany.
E-mail: andreas.eichhoefer@kit.edu; Tel: +49-(0)721-608-26371
cLehn Institute of Functional Materials, Sun Yat-Sen University, Guangzhou 510275,
China
dKarlsruhe Nano Micro Facility (KNMF), Hermann-von-Helmholtz-Platz 1, 76344
Eggenstein-Leopoldshafen, Germany
that ternary metal chalcogenide and chalcogenolato complexes
which comprise additional tin atoms in the copper chalcogen
framework exhibit interesting luminescent properties in the
solid state. The heteroheptanuclear Sn(IV)–Cu(I) oxosulphur
complex [Sn3Cu4(S2C2H4)6(μ3-O)(PPh3)4](ClO4)2 generates, for
example, a blue-green light at 495 nm,14 whereas [Cu2SnS
(edt)2(PPh3)2] (edt = (SCH2CH2S)2−) and [Cu2Sn(SPh)6(PPh3)2]
emit at 568 and 696 nm, respectively.15 The large cluster
complex (NBu4)[Cu19S28(SnPh)12(PEt2Ph)3] was found to
display deep-red, temperature-dependent photoluminescence
at ∼820–930 nm, which is especially bright at temperatures
below ∼100 K.16 Despite the great diversity of luminescence
†Electronic supplementary information (ESI) available: Molecular structures,
crystallographic data, X-ray powder patterns, calculated singlet excitation ener-
gies, additional optical spectra. CCDC 1487085 and 1487086. For ESI and crystal-
lographic data in CIF or other electronic format see DOI: 10.1039/c6dt04287b
This journal is © The Royal Society of Chemistry 2017
Dalton Trans.

View Article Online
Paper
Dalton Transactions
phosphorus atoms (Scheme 1 and Fig. S1, ESI†). The mole-
cular structure of 1 can be viewed as consisting of a distorted
octahedrally coordinated tin atom sharing parallel trigonal
faces with two copper centred ‘CuSe3P’ tetrahedra. This struc-
tural motif can also be found in complexes 2 (Table S1 and
Fig. S2, ESI†), 3 15 and 4 18 although with differing geometrical
parameters due to an exchange of the metal and chalcogen
atoms. Assuming the phenyl-chalcogenolato ligands to be
singly negatively charged we assign, in agreement with pre-
vious papers on 3 and 4, the formal oxidation states +IV to the
tin atoms and +I to the coinage metal atoms in 1 and 2. For a
comparison of the structural parameters of 1–4 see Table S2,
ESI.† In the literature, several other heterobimetallic trinuclear
complexes can be found, which comprise a similar linear
metal skeleton, including [Cu2Nb(SePh)6(PR3)2] (R = org.
Scheme 1 Structures of the trinuclear metal chalcogenolato com-
plexes 1–4.
properties found for copper-based complexes, to our know-
ledge, no examples were found with the NIR emission beyond
∼950 nm – in the spectral region which is of interest, for
instance, for bioanalytical applications. In fact, such emitters,
especially those with relatively high PL quantum yields, are
extremely rare among other metal complexes as well (NIR
emission is ubiquitous for lanthanide complexes, but is
usually of poor efficiency).17
group),19 [Co2V(SPh)8]2−, [Ni2V(SPh)8]2− 20
[Cu2M(SAr)6(PPh3)2]
,
(M = W, Mo, U; Ar = Ph, p-C6H4Me, p-C6H4F, p-C6H4Cl,
p-C6H4Br),21–23 and [LFeMFeL]n+ (M = Ge, Sn; n = 2, 3; L =
1,4,7-(4-tert-butyl-2-mercaptobenzyl)-1,4,7-
triazacyclononane).24
In the following we report on the four isostructural trinuc-
lear complexes [M2M′(EPh)6(PPh3)2] 1–4 (Scheme 1), which
luminesce at as long wavelengths as ∼900–1100 nm. We
describe in detail the synthesis of 1 and 2, crystallographic
data for 1–4, their absorption and emission spectra, as well as
the results of modelling these spectra via time-dependent
density functional theory (TDDFT) calculations.
Absorption spectra
Absorption spectra of 1–4 have been measured for powdered
crystals in a mineral oil layer between two quartz plates and in
solution (Fig. 1, 2 and S3, ESI†). Additionally, singlet transition
energies were calculated using the experimental structure para-
meters within time-dependent density functional theory
(TDDFT). Calculations were done with TURBOMOLE25,26
employing Becke’s three parameter hybrid functional with
Lee–Yang–Parr correlation (B3LYP),27,28 and polarized double-
zeta valence basis sets def2-SV(P),29 throughout. In Table 1
(compd 1 and 4) and Table S3† (compd 2 and 3), the calculated
energies for excitations with significant oscillator strengths
Results and discussion
Syntheses and structures
The synthesis and structural data have already been published
for 3 15 and 4.18 From reaction solutions of CuCl and
[SnCl4(thf)2] with PhSeSiMe3 in the presence of PPh3, 1 can be
obtained in the form of dark red crystals according to
Scheme 2. Yellow crystals of 2 can be obtained from a similar
reaction of AgCl, [SnCl4(thf)2], PhSSiMe3 and PPh3.
Single crystal X-ray analysis reveals that compound 1 crystal-
̅
lizes in the trigonal space group R3 (Table S1†). There is an
inversion centre at the tin atom and a threefold rotation axis
runs through the linear chain formed by the copper, tin and
Fig. 1 Comparison of measured electronic spectra (powdered crystals
in mineral oil) of 1 with calculated singlet excitation energies and oscil-
lator strengths plotted as vertical lines (green) as well as with super-
imposed Gaussians of FWHM = 0.3 eV (black curve) to simulate the
spectrum (see also Table 1).
Scheme 2 Synthesis of the trinuclear metal chalcogenolato complexes
1 and 2.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2017

View Article Online
Dalton Transactions
Paper
between the calculated singlet excitations and measured solid
state spectra of 1–4.‡
Note that the solid state spectra overemphasize the inten-
sity of weak bands due to the non-validity of Lambert–Beer law
for this kind of sample preparation. This is especially well
seen for the first absorption maximum of 4 at 1.43 eV, whereas
the relative band intensities in the spectrum of 4 in toluene
display a better agreement with the calculated excitations. For
complexes 1–3, absorption measurements in solution were not
possible due to decomposition as indicated either by strong
shifts of the absorption maxima of 1 and 3 (Fig. S4, ESI†) or by
formation of a white precipitate by 2. Respective characteristics
of excitations listed in Table 1 (Table S3, ESI†) are very similar
for compounds 1–3, which all exhibit S6 symmetry: they all are
transitions from HOMO−1/HOMO−2/HOMO−5 to the LUMO.
The reason for this similarity is found in the similarity of the
MO schemes of compounds 1–3 (Fig. 3 and S5, ESI†). The
LUMO is energetically well separated from the higher lying
unoccupied MOs (by more than 1 eV) and thus is the only
unoccupied orbital in reach for the lowest excitations. It is of
ag symmetry, therefore only transitions from occupied orbitals
of au or eu symmetry are dipole-allowed. For compounds 1–3,
these are the three orbitals mentioned above. One might note
that for instance for 1 the energetically lowest excitation corres-
ponds to a transition from HOMO−2 to LUMO, whereas that
from HOMO−1 to LUMO is slightly higher in energy. This can
happen, as excitations in TDDFT are calculated as the response
of the electron density on a time-dependent external electric
field. Thus not always an excitation is dominated by a pair of a
Fig. 2 Comparison of measured electronic spectra (powdered crystals
in mineral oil and in toluene) of 4 with calculated singlet excitation ener-
gies and oscillator strengths plotted as vertical lines (green) as well as
with superimposed Gaussians of FWHM = 0.3 eV (black curve) to simu-
late the spectrum (see also Table 1).
Table 1 Observed and calculated (TDDFT) electronic excitation ener-
gies in 1 and 4 a,b
Theory
Exp.
ΔE
ΔE
f
Character
1
4
2.39
2.60
3.52
0.0050
0.1852
0.2688
eu(HOMO−2) → ag(LUMO)
au(HOMO−1) → ag(LUMO)
eu(HOMO−5) → ag(LUMO)
2.57
3.65
1.43
1.96
2.43
1.59
2.01
2.6 × 10−5
0.0086
a(HOMO) → e(LUMO)
0.5 e(HOMO) → e(LUMO+1)
0.5 a(HOMO−1) → a(LUMO)
0.5 e(HOMO) → e(LUMO+1)
0.5 a(HOMO−1) → a(LUMO)
e(HOMO−2) → e(LUMO)
e(HOMO−3) → e(LUMO)
e(HOMO−5) → e(LUMO)
2.19
0.0757
2.24
2.57
2.70
0.0075
0.0489
0.0142
a Parameters: transition energy ΔE [eV], corresponding oscillator
strength f. For the spectra, see Fig. 1 and 2. b In the case of e represen-
tations the notation of the HOMOs/LUMOs counts for two degenerate
orbitals.
Fig. 3 Calculated molecular orbital diagrams of 1 and 4. The dashed
line separates occupied orbitals from unoccupied ones (eg/eu denotes
two levels, thus four near-degenerate orbitals).
(>5 × 10−3) are listed together with the respective molecular
orbitals involved and the corresponding values determined in
the experiment. From these transitions the spectra were simu-
lated by superimposing Gaussians with a full width at
half maximum of (FWHM) 0.3 eV. Fig. 1, 2 and Tables 1, S3
‡Excitation spectra of 1–4 for the DFT-optimized structures qualitatively agree
with those for the experimental structures, but are shifted to lower energies by
ca. 0.5 eV. They are shown in the supplement together with the relevant bond
distances (Fig. S13 and Table S9, ESI†). Respective optimized Cartesian coordi-
(ESI†) demonstrate a very good agreement within ∼0.2 eV nates are available (ESI†).
This journal is © The Royal Society of Chemistry 2017
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Fig. 4 Molecular frontier orbitals involved in the relevant (lowest
energy) electronic transitions of 1. Contours are drawn at 0.05 (white)
and −0.05 (black) atomic units [a.u.].
Fig. 5 Molecular frontier orbitals involved in the relevant (lowest
energy) electronic transitions of 4. Contours are drawn at 0.05 (white)
single occupied and a single virtual molecular orbital, and
even if this is the case, the difference of the respective orbital and −0.05 (black) atomic units [a.u.].
energies is only a rough approximation to the excitation
energy. As a consequence, the sequence of excitations (Tables
1 and S3, ESI†) is not necessarily the same as the sequence of
orbital energy differences depicted in Fig. 3 and S5 in the
ESI.† Visualization of these orbitals, as shown in Fig. 4 exem-
plarily for compound 1, reveals that the LUMO is an anti-
bonding combination of the 5s orbital of the central tin atom
and the 4p orbitals of the neighboring sulfur atoms, whereas
the occupied orbitals do not involve the tin atom but copper d,
sulfur p and – for HOMO−1 – also phosphorus p orbitals. For
compound 4 (C3 symmetry) matters are somewhat different.
LUMO (of e symmetry) and LUMO+1 (of a symmetry) are quite
close in energy (∼0.4 eV, see Fig. 3) and thus both are involved
in the transitions listed in Table 1. In contrast to 1–3 they
show contributions predominantly of the d orbitals of the
central titanium atom and almost no contributions of the
neighboring sulphur atoms (Fig. 5). Furthermore, all occupied
orbitals of high energy are involved in the transitions, as no
symmetry applies.
Emission spectra
Solid (crystalline) complexes 1–4 display relatively strong NIR
emission centered between 920 and 1110 nm at low tempera-
tures (Fig. 6, Table 2, and Fig. S6, S7, ESI†). However, the PL
intensity of 1–3 decreases significantly by increasing the temp-
erature over ∼100 K. In these compounds the PL decays on the
Fig. 6 Photoluminescence excitation (PLE, solid line) and emission (PL,
dashed line) spectra of 1 and 4 (powdered crystals in mineral oil) at
time scale of sub-microseconds to tens of microseconds as
extracted from bi-exponential fits. Consequently, the PL can be different temperatures.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2017

View Article Online
Dalton Transactions
Paper
Table 2 Photoluminescence wavelength λPL [nm] at 150 and 20 K and
decay times τ [μs] for 1–4
parameters were optimized for the respective excited states,
accounting for geometric relaxation after excitation. Finally,
emission energies for the respective excited state were
obtained as excitation energies from a ground-state determi-
nant calculated for the relaxed structure (Table S8, ESI†). For
1–3 this method was applied to the four lowest triplet tran-
sitions and for 4 to the two lowest singlet transitions. The cal-
culations predict large Stokes shifts of 0.72–1.55 eV for 1–3
and small ones (0.09–0.27 eV) for 4. Both trends are in fact
also observed in the experiments: Stokes shifts of ca. 1.4–1.6
eV are estimated for 1–3, whereas for 4 the shift is ca. 0.3 eV.§
This difference in Stokes shifts between compounds 1–3
(with tin in the central position) and 4 (with titanium in the
λPL, 150 K
λPL, 20 K
τ, 150 K a
τ, 20 K a
1
1090, 715
1110, 600
0.59, 0.17 (34 : 66)
54, 8.4 (50 : 50)
1.7, 0.32 (49 : 51)
63, 7.7 (60 : 40)
35, 7.9 (55 : 45)
10, 1.5 (76 : 24)
115, 25 (40 : 60)
0.04
2
3
930
1020, 700
930
1020, 685
4
1090b
1063
∼0.01b
a Biexponential decay parameters for 1–3 with relative amplitudes of
two exponential components in brackets. b Parameters at 295 K.
attributed to phosphorescence. In contrast to 1–3, complex 4 central position) can be related to the emission character
remains an efficient emitter also at ambient temperature, with (phosphorescence vs. fluorescence for 1–3 and 4, respectively),
the emission maximum at ∼1090 nm and a relatively high but also to the degree of the distortion in the excited state
quantum efficiency of 9.5% – comparable to that of NIR-emit- (Table S9, ESI†), in particular to the changes in distances
ting quantum dot nanoparticles.30
between the central atom and its neighbors upon electronic
Referring to the last value, the PL efficiency of ∼30% is esti- excitation. For 4, only comparably small structure relaxation
mated for 4 below ∼50 K. The NIR emission of 4 proceeds un- effects upon electronic excitation are observed: the Ti–S bonds
usually fast, with a lifetime of ∼40 ns at 20 K, which decreases are elongated by ca. 5 pm. Much more pronounced elongation
to ∼10 ns at ambient temperature – close to the time resolu- upon excitation is calculated for the Sn–E (E = S, Se) distances
tion limit of our apparatus. Consequently, the PL of 4 can be in 1–3: 12–20 pm, depending on the elements and the number
ascribed to fluorescence. In addition the following details of of the excited state. These observations might be rationalized
the experimental luminescence spectra are noteworthy. with the character of the LUMO (Fig. 4), which is occupied for
Although PL excitation (PLE) spectra (Fig. 6, S5 and S7, ESI†) all excited states discussed. For the excited state structures of
agree well with the corresponding absorption spectra (Fig. 1 1–3 it is anti-bonding with respect to the central atom and its
and 2) with regard to the onsets of the lowest-energy bands, neighbors, so distances will increase if it is occupied. On the
the ‘saturation’ effect for PLE features is even more pro- other hand, for elongated distances, the anti-bonding charac-
nounced than that in the solid state absorption (see above) ter of respective orbitals is reduced, which in turn lowers the
due to optically ‘thick’ sample preparations for PL measure- orbital energy and thus also the energy of the respective
ments. We also remark that the NIR emission of 4 at excited state resulting in ΔEem ≪ ΔEex.¶ For 4, in contrast, the
∼1150 nm was also detected in toluene solution (Fig. S8, ESI†), LUMO (Fig. 5) is essentially non-bonding (3d orbitals of tita-
however with a much weaker intensity as from the solid nium), which leads both to only moderate deviations in the
samples. We note in passing that 3 demonstrates a second structural parameters upon excitation and to smaller changes
minor emission band at 685 nm at low temperatures. A dual in the orbital energies. Such large structural relaxations
emission is also observed for 1 under UV excitation (Fig. S9, accompanied by energy lowering for excited states are not
ESI†). The higher-energy emission bands decay significantly uncommon. For instance, Costuas et al.31 calculated an
slower as compared to the NIR bands (Table 2). Combined elongation of Mo–Mo distances for the first excited triplet state
theoretical and time-resolved photoluminescence investi- of [Mo6Br14]2− by ca. 12 pm, accompanied by an energetic
gations of [Mo6Br14]2− metal cluster units have recently evi- lowering of 0.4 eV. As explicitly noted in their work, consider-
denced the existence of multiple competitive de-excitation pro- ation of these relaxation effects is essential for the match of
cesses in these compounds – in contrast to a ‘simple’ relax- the calculated and measured emission energies.
ation pathway as described by Kasha’s rule.31 The origin of the
dual emission in these complexes needs to be clarified in
future work.
§Stokes shifts were estimated for 1–3 and 4 from the PLE and PL spectra at low
One general discrepancy evident by comparison of the
absorption and emission spectra of 1–4 is the distinctly
different Stokes shifts of 1–3 versus 4. In order to identify poss-
and ambient temperatures, respectively. At cryogenic temperatures, the Stokes
shift of 4 slightly decreases to ca. 0.27 eV. Note also that the low-energy PLE
bands of 1, 2 (recorded at 17–20 K) and 4 (both at 20 and 295 K), well correspond
ible reasons for this behaviour, we made an attempt to simu-
to their counterparts in the Nujol absorption spectra measured at ambient temp-
late the respective processes with TDDFT – similar to the route erature; solid 3 shows a ∼0.25 eV upshift of the absorption at 295 K vs. the PLE
followed by Costuas et al.31 – in the following manner. At first,
the structure parameters of 1–4 were optimized. For the result-
onset at 17 K (see Fig. 1, 2, 6, S6 and S7†).
¶According to a Mulliken population analysis for the spin density of the excited
triplet state in 1–3, the unpaired electrons are mainly located at the chalcogenide
ing structures triplet and singlet excitation energies were calcu-
atoms, 0.20 to 0.31 electrons per atom (R. S. Mulliken, Electronic Population
lated for 1–3 and 4, respectively, in order to simulate the
Analysis on LCAO–MO Molecular Wave Functions. I. J. Chem. Phys., 1955, 23(10),
respective excitation process (Table S8, ESI†). Next, structure 1833–1840).
This journal is © The Royal Society of Chemistry 2017
Dalton Trans.

View Article Online
Paper
Dalton Transactions
singlet transitions in 4, both in excitation and emission, are
predicted with low oscillator strength. This is consistent with
the weak red-NIR absorption of 4, but in discrepancy with its
efficient NIR fluorescence, which remains to be clarified by
future investigations.
Experimental section
Synthesis
Scheme 3 Schematic representation of the electronic processes in 1–3
(left) and 4 (right), representatively shown for the energetically lowest
transition (ISC ≡ intersystem crossing). ΔEexc − ΔEem depicts the calcu-
lated Stokes shift listed in Table S7, ESI.†
Standard Schlenk techniques were employed throughout the
syntheses using a double manifold vacuum line with high
purity dry nitrogen (99.9994%) and a MBraun glovebox with
high purity dry argon (99.9990%). The solvents Et2O (diethyl-
ether), thf (tetrahydrofuran) and toluene were dried over
sodium-benzophenone, and distilled under nitrogen.
A schematic overview of the electronic processes in 1–4 is
shown in Scheme 3. For 1–3 the molecules are excited from
the ground state S0 to the lowest excited singlet state S1*
(ΔEexc) which then converts by intersystem crossing to an
excited triplet state T*.
Distinct geometric relaxation of this excited state structure
results in the relaxed excited triplet state T from which, at low
temperature, phosphorescence (ΔEem) with a strong Stokes
shift (ΔEexc − ΔEem) can be observed. In 4 there is no inter-
system crossing to a triplet state and only a moderate
relaxation of the excited state structure S1* to S1 comprising a
fluorescence with a small Stokes shift.
We note that calculations for 1–4 were done in gas phase.
Distinct differences in bond lengths (about 10 pm) between
the optimized and experimentally determined ground state
structures to a certain degree indicate deficits of this simplifi-
cation compared to a treatment of the solid state. The latter
might give a more reliable picture but lies outside the scope of
present technologies.
Anhydrous dimethylenchlorine (CH2Cl2) (H2O
< 0.005%)
obtained from Aldrich was degassed, freshly distilled and
stored over molecular sieves under nitrogen. CuCl obtained
from Sigma-Aldrich was subsequently washed with HCl,
CH3OH, and diethylether to remove traces of CuCl2, and dried
under vacuum. [SnCl4(thf)2],32 [AgCl(PPh3)3]33 and LiSPh34
were prepared according to literature procedures. PPh3
obtained from Sigma-Aldrich was recrystallized in ethanol
before use.
PhSSiMe3. LiSPh (67.3 mmol, 7.82 g) was suspended in
200 mL of thf in a 1 L flask and Me3SiCl (7.39 g, 68 mmol),
diluted with 50 mL of thf, was added dropwise to the solution
upon stirring. The reaction mixture was then heated for 8 h
under reflux after which thf was removed by distillation and
then the product was removed from the remaining wet, solid
residue by vacuum distillation (1 mbar) into a flask connected
with a knee. This raw product (turbid liquid) was then further
purified by a vacuum distillation (8 mbar) with a vigreux
column to yield PhSSiMe3 as a colourless liquid (b.p. 70 °C).
1
Yield: 10.6 g (86.4%). C9H14SSi (182.36): H (500 MHz, C6D6):
7.38 (m, 2H, ortho-CH), 6.95 (m, 3H, meta, para-CH); 0.12(s,
3H, SiCH3); 13C{1H}-NMR (126 MHz, C6D6): 134.6 (s, ortho-CH),
Conclusions
We conclude that the experimental optical absorption spectra 131.2 (s, CSSi), 128.1 (s, meta-CH), 126.1 (s, para-CH) ppm, 0.0
of the trinuclear complexes studied in this work are well (s, SiCH3); 29Si (99 MHz, C6D6): 16.6 ppm.
described by the TDDFT calculations. This is in line with our
[Cu2Sn(SePh)6(PPh3)2] (1). [SnCl4(thf)2] (0.31 g, 0.77 mmol)
previous results that the TDDFT method can generally well and CuCl (0.152 g, 1.53 mmol) were dissolved with PPh3 (0.5 g,
describe the absorption spectra of various polynuclear copper(I) 1.91 mmol) in 30 mL of thf to give a light yellow solution.
complexes. At temperatures below ∼100 K, solid compounds Upon addition of PhSeSiMe3 (0.91 mL, 4.75 mmol) the colour
1–3 demonstrate bright near-infrared (NIR) microsecond-long of the solution turned quickly to orange red. After 3 h stirring
phosphorescence at ∼900–1100 nm. Its intensity decreases at room temperature thf was removed by vaccum condensation
strongly by increasing the temperature. Complex 4 also shows and the residue dissolved in 15 mL of CH2Cl2. After addition
NIR emission at ∼1100 nm, but with an unusually high, practi- of 10 ml of Et2O dark red crystals of 1 started to grow after one
cally relevant efficiency of 9.5% at ambient temperature com- night in the fridge at −42 °C. After four days the crystals were
parable to that of NIR-emitting quantum dot nanoparticles. filtered cold and were washed three times with 10 mL of Et2O
The decay is fast (<10 ns), i.e. the emission is fluorescence. to give a final yield of 1.24 g (80.0%). C72H60Cu2P2Se6Sn
The trends in the experimental Stokes shifts of 1–4 correlate (1706.76): calcd C 50.7, H 3.5 found C 50.5, H 3.8%.
with the amount/size of the calculated geometrical relaxation
[Ag2Sn(SPh)6(PPh3)2] (2). [SnCl4(thf)2] (0.31 g, 0.77 mmol)
in the respective excited states (i.e. lowering of the orbital ener- and [AgCl(PPh3)3] (1.42 g, 1.52 mmol) were suspended in
gies) and can be rationalized by the qualitative difference of 30 mL of thf. Upon addition of PhSSiMe3 (0.91 mL,
the LUMOs of 1–3 compared to 4. However, the lowest-energy 4.75 mmol) the colour of the solution turned to pale yellow
Dalton Trans.
This journal is © The Royal Society of Chemistry 2017

View Article Online
Dalton Transactions
Paper
but remains turbid. After 3 h stirring at room temperature thf program package finding a total of 75 (1) and 120 (2) electrons
was removed by vacuum condensation and the residue dis- in a potential solvent accessible area of 471 (1) and 445 (2) Å.
solved in 15 mL of CH2Cl2 to give a clear yellow solution. This
CCDC 1487085 (1) and 1487086 (2) contain the supplemen-
was ‘layered’ with diethylether through slow diffusion by tary crystallographic data for this paper.
evaporation via a connected flask to give pale yellow crystals of
X-ray powder diffraction patterns (XRD) for 1–3 were
2 which were filtered after four days and washed three times measured on a STOE STADI P diffractometer (Cu–Kα1 radi-
with 10 mL of Et2O to give a final yield of 0.82 g (71.3%). ation, Germanium monochromator, Debye–Scherrer geometry,
C72H60Ag2P2S6Sn (1514.03): calcd C 57.1, H 4, S 12.7 found Mythen 1K detector) in sealed glass capillaries both as a sus-
C 56.4, H 4.5, S 12%. The reduced experimental values for pension of crystals in the mother liquor and as a powder of
C and S (enhanced for H) are indicative for remaining lattice crystals. The theoretical powder diffraction patterns were cal-
solvent molecules of CH2Cl2 and (C2H5)2O. The single crystal culated on the basis of the atom coordinates obtained from
X-ray analysis also reveals the exsistence of solvent accessible single crystal X-ray analysis by using the program package
voids although no such molecules could be properly refined STOE WinXPOW.39 A comparison of the measured and calcu-
due to disorder.
lated X-ray powder diffraction patterns for 1–3 reveals the crys-
[Cu2Sn(SPh)6(PPh3)2] (3). SnCl4(thf)2 (0.31 g, 0.77 mmol) talline purity with respect to the formation of other crystalline
and CuCl (0.152 g, 1.53 mmol) were dissolved with PPh3 (0.5 g, compounds (Fig. S10–S12, ESI†). Slightly increasing differ-
1.91 mmol) in 30 mL of CH2Cl2 to give a pale yellow solution. ences in the position of the peaks with increasing detection
Upon addition of PhSSiMe3 (0.91 mL, 4.75 mmol) the colour angle arise from the temperature difference of the data collec-
of the solution intensified and went to orange. Stirring was tions (single crystal XRD at 180 K and powder XRD at room
stopped and crystallization of 3 started. After four days the temperature).
supernatant solution was decanted and the crystals washed
Physical measurements
three times with 10 mL of thf to give a final yield of 0.88 g
(80.0%). C72H60Cu2P2S6Sn (1425.39): calcd C 60.7, H 4.2, S 13.5
found C 61.2, H 4.1, S 12.7%.
C, H, S elemental analyses were performed on an ‘Elementar
vario Micro cube’ instrument.
[Cu2Ti(SPh)6(PPh3)2] (4) was prepared according to ref. 18.
UV-vis absorption spectra were measured on a Perkin Elmer
Lambda 900 spectrophotometer. Solid state spectra were
measured in transmission for samples which were prepared as
micron-sized crystalline powders dispersed in a nujol layer
between two quartz plates. These were placed in front of a
Labsphere integrating sphere.
Photoluminescence (PL) measurements were performed on
a Horiba JobinYvon Fluorolog-3 spectrometer equipped with a
Hamamatsu R5509 vis-NIR photomultiplier (∼300–1400 nm)
and an optical close-cycle cryostat (Leybold) for measurements
at cryogenic temperatures down to ca. 17 K. The emission
spectra were corrected for the wavelength-dependent response
of the spectrometer and detector (in relative photon flux
units). Emission decay traces were recorded by connecting the
photomultiplier to an oscilloscope (typically with a 500 or 50
Ohm load) and using a N2-laser for pulsed excitation at
Crystallography
Crystals suitable for single crystal X-ray diffraction were
selected in perfluoroalkylether oil and mounted on a diffracto-
meter equipped with an Oxford Cryosystem. Single-crystal
X-ray diffraction data of 1 were collected using graphite-mono-
chromatised MoKα radiation (λ = 0.71073 Å) on a STOE IPDS
2T (Imaging Plate Diffraction System). Single-crystal X-ray dif-
fraction data of 2 were collected using MoKα radiation (λ =
0.71073 Å) generated by a multilayer optic on a STOE STADI
Vari (Pilatus Hybrid Pixel Detector 300K). Raw intensity data
were collected and treated with the STOE X-Area software
Version 1.64. Interframe Scaling of the STADI Vari dataset of
2 was done with the implemented program LANA. Data were cor-
rected for Lorentz and polarization effects. Based on a crystal
description a numerical absorption correction was applied.35
All structures were solved with the direct methods program
SHELXS of the SHELXTL PC suite programs,36 and were
refined with the use of the full-matrix least-squares program
SHELXL. Molecular diagrams were prepared using Diamond.37
In 1 and 2 all Cu, Ag, S, Se, P and C atoms were refined with
anisotropic displacement parameters whilst H atoms were
computed and refined, using a riding model, with an isotropic
temperature factor equal to 1.2 times the equivalent tempera-
ture factor of the atom which they are linked to. In 2 phenyl
rings of the phosphine ligands are disordered and C atoms
were refined with a split model of site disorder. Lattice solvent
molecules were identified within both structures. These were
337 nm (∼2 ns, ∼5 μJ per pulse). PL quantum yield of 4, φPL
,
was measured at ambient temperature using 10 cm
a
integrating sphere out of optical PTFE with low auto-lumine-
scence (Berghof GmbH), which was installed in the sample
chamber of the Fluorolog-3. A powdered sample of 4 between
quartz plates (see above) was placed into the sphere
and excited at 500 nm. The φPL value was determined
according to the method of de Mello et al.40 The accuracy was
estimated to be 10%.
Author contributions
located on a threefold axis and could not be adequately A. E. was responsible for synthesis and characterization,
refined due to disorder. The data were therefore corrected for S. L. for PL/PLE measurements, M. K. and F. W. for DFT and
these using the SQUEEZE option within the PLATON38 TDDFT calculations.
This journal is © The Royal Society of Chemistry 2017
Dalton Trans.

View Article Online
Paper
Dalton Transactions
17 J.-C. Bünzli, Coord. Chem. Rev., 2015, 293–294, 19–47.
Conflict of interests
18 V. Andrushko, H. Sommer, D. Himmel, D. Fenske and
A. Eichhöfer, Eur. J. Inorg. Chem., 2011, 3102–3110.
19 B. Bechlars, R. Feuerhake and D. Fenske, Z. Anorg. Allg.
Chem., 2007, 633, 2603–2613.
The authors declare no financial interests.
20 A. Eichhöfer, T. Bodenstein and K. Fink, Eur. J. Inorg.
Chem., 2014, 3510–3520.
Acknowledgements
This work was supported by the Karlsruhe Institute of
Technology (KIT, Campus North). The authors wish to thank
A. K. Powell for generous support and S. Stahl for performing
the elemental analysis.
21 J. M. Ball, P. M. Boorman, J. F. Fait and T. J. Ziegler,
J. Chem. Soc., Chem. Commun., 1989, 722–723.
22 P. C. Leverd, M. Lance, M. Nierlich, J. Vigner and
M. Ephritikhine, J. Chem. Soc., Dalton Trans., 1994, 3563–
3567.
23 P. M. Boorman, H. B. Kraatz, M. Parvez and T. Ziegler,
J. Chem. Soc., Dalton Trans., 1993, 433–439.
24 T. Glaser, E. Bill, T. Weyhermuller, W. Meyer-Klaucke and
K. Wieghardt, Inorg. Chem., 1999, 38, 2632–2642.
25 TURBOMOLE Version 7.0, TURBOMOLE GmbH, 2015.
Notes and references
1 V. W.-W. Yam, K. K.-W. Lo, W. K.-M. Fung and C.-R. Wang,
Coord. Chem. Rev., 1998, 171, 17–41.
TURBOMOLE is
a development of the University of
2 P. C. Ford and A. Vogler, Acc. Chem. Res., 1993, 26, 220–226.
3 L. Bergmann, G. J. Hedley, T. Baumann, S. Bräse and
D. W. Samuel, Sci. Adv., 2016, 2, e1500889.
4 T. Hofbeck, U. Monkowius and H. J. Yersin, J. Am. Chem.
Soc., 2015, 137, 399–404.
5 M. Knorr, A. Bonnot, A. Lapprand, A. Khatyr,
C. Strohmann, M. Kubicki, Y. Rousselin and P. D. Harvey,
Inorg. Chem., 2015, 54(8), 4076–4093.
Karlsruhe and the Forschungszentrum Karlsruhe
1989–2007, TURBOMOLE GmbH since 2007.
26 F. Furche, R. Ahlrichs, C. Hättig, W. Klopper, M. Sierka and
F. Weigend, Wiley Interdiscip. Rev.: Comput. Mol. Sci., 2014,
4, 91–100.
27 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens.
Matter, 1988, 37, 785–789.
28 A. D. J. Becke, Chem. Phys., 1993, 98, 5648–5652.
29 F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005,
7, 3297–3305.
30 Q. Ma and X. Su, Analyst, 2010, 135, 1867–1877.
31 K. Costuas, A. Garreau, B. Bulou, B. Fontaine, J. Cuny,
R. Gautier, M. Mortier, Y. Molard, J.-L. Duvail, E. Faulques
and S. Cordier, Phys. Chem. Chem. Phys., 2015, 17, 28574–
28585.
32 S. J. Ruzicka and A. E. Merbach, Inorg. Chim. Acta, 1976, 20,
221.
33 D. V. Sanghani, P. J. Smith, D. W. Allen and B. F. Taylor,
Inorg. Chim. Acta, 1982, 59, 203–206.
6 R. Langer, M. Yadav, B. Weinert, D. Fenske and O. Fuhr,
Eur. J. Inorg. Chem., 2013, 3623–3631.
7 C. Yue, C. Yan, R. Feng, M. Wu, L. Chen, M. Jiang and
M. Hong, Inorg. Chem., 2009, 48, 2873–2879.
8 L. Bergmann, J. Friedrichs, M. Mydlak, T. Baumann,
M. Nieger and S. Bräse, Chem. Commun., 2013, 49, 6501–
6503.
9 S. Igawa, M. Hashimoto, I. Kawata, M. Yashima, M. Hoshino
and M. Osawa, J. Mater. Chem. C, 2013, 1, 542–551.
10 Y.-M. Wang, F. Teng, Y.-B. Hou, Z. Xu, Y.-S. Wang and
W.-F. Fu, Appl. Phys. Lett., 2005, 87, 233512.
11 E. Cariati, E. Lucenti, C. Botta, U. Giovanella, D. Marinotto
and S. Righetto, Coord. Chem. Rev., 2016, 306, 566–614.
12 X.-X. Yang, I. Issac, S. Lebedkin, M. Kühn, F. Weigend,
D. Fenske, O. Fuhr and A. Eichhöfer, Chem. Commun.,
2014, 50, 11043–11045.
34 R. Köster, G. Seidel and R. Boese, Chem. Ber., 1988, 121,
1137–1142.
35 X-RED32 1.01, Data Reduction Program, STOE&CIE GmbH,
Darmstadt, Germany, 2001.
36 G. M. Sheldrick, SHELXTL 6.14, An Integrated System for
Solving, Refining, and Displaying Crystal Structures from
Diffraction Data, Bruker Analytical X-ray Systems,
BrukerAXS Inc., Madison Wisconsin, 2003.
37 K. Brandenburg, Diamond Version 2.1d, Crystal Impact
GbR, 1996–2000.
13 A. Eichhöfer, G. Buth, M. Kühn and F. Weigend, Inorg.
Chem., 2015, 54, 9413–9422.
14 X. Wang, T.-L. Sheng, R.-B. Fu, S.-M. Hu, S.-C. Xiang,
L.-S. Wang and X.-T. Wu, Inorg. Chem., 2006, 45, 5236–
5238.
15 L.-S. Wang, T.-L. Sheng, X. Wang, D.-B. Chen, S.-M. Hu,
R.-B. Fu, S.-C. Xiang and X.-T. Wu, Inorg. Chem., 2008, 47,
4054–4059.
38 A. L. Spek, Acta Crystallogr., Sect. A: Fundam. Crystallogr.,
1990, 46, C-34.
39 STOE, WinXPOW, STOE & Cie GmbH, Darmstadt, 2000.
40 J. C. de Mello, H. F. Wittmann and R. H. Friend, Adv.
Mater., 1997, 9, 230.
16 A. Eichhöfer, J. Jiang, S. Lebedkin, D. Fenske,
D. G. McDonald, J. F. Corrigan and C.-Y. Su, Dalton Trans.,
2012, 41, 3321–3327.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2017