Beyond Classical Coordination Chemistry: The First Case of a Triply Bridging Phosphine Ligand

Article abstract of DOI:10.1002/anie.202103037

The first example of a triply bridging (μ₃-P) phosphine ligand has been discovered in the crown-shaped [Cu₃(μ₂-Hal)₃L] (Hal=Cl, Br, or I) complexes supported by tris[2-(2-pyridyl)ethyl]phosphine (L). Theoretical analysis completely confirms the observed μ₃-P-bridging pattern, revealing the interaction of the same lone pair of phosphorus with three valence 4s-orbitals of Cu atoms. The presented complexes exhibit outstanding blue phosphorescence (λ_em=442–465 nm) with the quantum efficiency reaching 100 %. The complex [Cu₃(μ₂-I)₃L] also exhibits remarkable thermo- and mechanochromic luminescence resulting in a sharp change in the emission colour upon external stimuli. These findings essentially contribute to coordination chemistry of the pnictine ligands.

Full text of DOI:10.1002/anie.202103037

Angewandte
Research Articles
Chemie
How to cite:
International Edition: doi.org/10.1002/anie.202103037
German Edition: doi.org/10.1002/ange.202103037
Coordination Modes
Beyond Classical Coordination Chemistry: The First Case of a Triply
Bridging Phosphine Ligand
Andrey Y. Baranov, Elena A. Pritchina, Alexey S. Berezin, Denis G. Samsonenko,
Vladimir P. Fedin, Nataliya A. Belogorlova, Nina P. Gritsan,* and Alexander V. Artem’ev*
Abstract: The first example of a triply bridging (m3-P)
phosphine ligand has been discovered in the crown-shaped
[Cu3(m2-Hal)3L] (Hal = Cl, Br, or I) complexes supported by
tris[2-(2-pyridyl)ethyl]phosphine (L). Theoretical analysis
completely confirms the observed m3-P-bridging pattern,
revealing the interaction of the same lone pair of phosphorus
with three valence 4s-orbitals of Cu atoms. The presented
were reported.[5–11] Around the same time, Reau et al.
disclosed a dipalladium(I) complex decorated by 2,5-bis(2-
pyridyl)phosphole ligands with m2-bridging phosphorus
atoms.[12] Afterwards, a family of dicopper(I) complexes with
the m2-phosphine coordination mode was prepared utilizing
such 2,5-bis(2-pyridiyl)phosphole ligands.[13–18] However, de-
spite these findings, m2-bridging pnictine ligands remain very
limited. At the moment, some 110 compounds have been
reported (most of them by Werner and Reau).[19]
Even more exotic is the m3-bridging mode of R3Pn ligands,
which was previously found only in four complexes. The first
one, a Pd3 triangle capped by triply bridging PF3 ligand,[20,21]
was reported in 1990 by Balch et al. (I, Figure 1). It should be
emphasized that the PF3 is not a typical phosphine ligand
complexes exhibit outstanding blue phosphorescence (lem
=
442–465 nm) with the quantum efficiency reaching 100%.
The complex [Cu3(m2-I)3L] also exhibits remarkable thermo-
and mechanochromic luminescence resulting in a sharp change
in the emission colour upon external stimuli. These findings
essentially contribute to coordination chemistry of the pnictine
ligands.
ꢀ
because of absence of C P bonds and its extremely p-
Introduction
acceptor character. Twenty years later, Gabbaꢀ et al. first
described triply coordinated heavier pnictines, which were
found in the Cu3 and Ag3 complexes stabilized by the [o-
(iPr2P)C6H4]3Pn ligands (Pn = Sb, Bi) [II, Figure 1].[22] The
most recent example, the [Pd4(m3-SbMe3)4(SbMe3)4] cluster
with four m3-SbMe3 ligands, was synthesized in 2016 by Reid
et al. (III, Figure 1).[23]
Nowadays, phosphines and their heavier pnictine ana-
logues, R3Pn (Pn = P, As, Sb or Bi), are widely used as ligands
in almost all fields of chemistry and materials science. The
highly important role of these ligands (foremost, phosphines)
in coordination chemistry, transition metal catalysis, and
materials science can hardly be overestimated.[1–3] For a long
time, it had been generally accepted that the R3Pn ligands can
act exclusively as terminal donors, binding a single metal
center.[1–3] This paradigm was first challenged in 1994 by
Werner et al., who discovered the m2-bridging mode of the
iPr3Sb ligand in the [Rh2Cl2(m-SbiPr3)(m-CR2)2] complexes.[4]
A few years after this breakthrough, the first examples of m2-
bridging R3P and R3As ligands, found in similar complexes,
[*] A. Y. Baranov, A. S. Berezin, Dr. D. G. Samsonenko, Prof. V. P. Fedin,
Dr. A. V. Artem’ev
Nikolaev Institute of Inorganic Chemistry, SB RAS
3 Acad. Lavrentiev Ave., 630090 Novosibirsk (Russian Federation)
E-mail: chemisufarm@yandex.ru
Dr. E. A. Pritchina, Prof. N. P. Gritsan
Voevodsky Institute of Chemical Kinetics and Combustion, SB RAS
3 Institutskaya Str., 630090 Novosibirsk (Russian Federation)
and
Novosibirsk State University
2, Pirogova Str., Novosibirsk 630090 (Russian Federation)
E-mail: gritsan@kinetics.nsc.ru
Dr. N. A. Belogorlova
A. E. Favorsky Irkutsk Institute of Chemistry, SB RAS
1 Favorsky Str., 664033 Irkutsk (Russian Federation)
Supporting information and the ORCID identification number(s) for
the author(s) of this article can be found under:
https://doi.org/10.1002/anie.202103037.
Figure 1. Known complexes containing m3-bridging pnictine(III) li-
gands.
&&&&
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 2 – 10
These are not the final page numbers!

Angewandte
Research Articles
Chemie
While the complexes with m2- and m3-bridging modes of
unexpectedly found that exploiting 3 equiv of the CuI
R3Pn ligands are still exotic in chemistry, they are immensely
important for understanding the nature of the bonding in the
[R3Pn-Mn] species (M is a metal, n = 2 or 3). The latter can be
much more numerous than now, for example, as transient
intermediates in catalytic reactions. Deep understanding of
these bonding patterns would facilitate the targeted design
and synthesis of the m2- and m3-bridging R3Pn ligands. Apart
from this, due to the unusual bonding pattern, the functional
properties of the m2- and m3-bridging pnictine compounds
might turn out to be far from trivial and indeed deserve
deeper scrutiny.
Herein we report on the first discovered m3-P-bridging
coordination of the phosphine ligand, disclosed in a series of
crown-shaped [Cu3(m2-Hal)3L] complexes (IV, Figure 1). The
latter were synthesized using tris[2-(2-pyridyl)ethyl]phos-
phine (L), where the phosphorus atom is m3-capped by
a supported Cu3 triangle. A complementary set of theoretical
techniques, viz., QTAIM, NBO, and natural energy decom-
position (NEDA) analyses performed on the top of DFT
calculations, confirmed the m3-bridging mode of the phos-
phine ligand and revealed a significant contribution of orbital
(or charge-transfer, CT) interactions to the bonding of the
phosphorus atom to the three CuI cations of the Cu3 core. It is
also worth mentioning that the new [Cu3(m2-Hal)3L] com-
plexes exhibit remarkable luminescent properties, such as
room temperature deep-blue phosphorescence with a quan-
tum efficiency close to 100%, nontrivial mechano- and
thermochromic luminescence along with the excitation-de-
pendent emission. These results essentially contribute to
coordination chemistry and the theory of chemical bonding.
Apart from this, from an application viewpoint, we proposed
a new way of designing luminophores with prominent stimuli-
responsive luminescence properties.
precursor in this reaction leads to structurally intriguing
[Cu3(m2-Hal)3L] complexes (Hal = Cl (1), Br (2), or I (3)),
which were isolated in 61–87% yields. According to single
crystal X-ray diffraction (sc-XRD) analysis, these products
show the pseudo-C3-symmetric structure (Figure 2). The
complexes consist of a crown-shaped Cu3(m2-Hal)3 cluster,
bridged at the corners by pyridine nitrogen atoms (dCu-N
ꢁ 2.00 ꢁ) and capped by the triply bridging phosphorus atom
of the phosphine ligand. Note that three Cu–P distances differ
significantly in the series. Namely, the two Cu–P bonds in the
chloride 1 and iodide 3 are approximately equal (av. 2.45 ꢁ
and 2.52 ꢁ), whereas the third one (av. 2.89 ꢁ) is noticeably
longer. In contrast, three Cu–P bonds in bromide 2, being
2.561, 2.523, and 2.632 ꢁ, are almost the same. Meanwhile,
given that all Cu–P distances of 1–3 are significantly shorter
than the sum of van der Waals (vdW) radii of Cu and P atoms
(3.2 ꢁ),[25] the m3-bridging pattern of the P atom takes place in
the entire series. According to the Cambridge structural
database, the coordinated terminal phosphines feature the
Cu–P distance being ca. 2.25 ꢁ, and the m2-bridging ones,
namely about 2.37 ꢁ. In turn, the Cu3 triangles of 1–3 are
similar in size with the Cu···Cu distances being about 2.48,
2.56, and 2.57 ꢁ. Such short contacts (twice vdW radius of Cu,
2.80 ꢁ[25]) clearly indicate rather strong metallophilic inter-
actions in the Cu3 unit. This fact is also confirmed theoret-
ically (see below).
The temperature dependent sc-XRD study of 3 (Support-
ing Information, Table S1) reveals that three Cu···Cu and Cu–
P distances vary nonlinearly upon a decrease of temperature
from 250 to 130 K (Supporting Information, Figures S1, S2).
Namely, in the range of 250–165 K, the Cu1···Cu2 and
Cu2···Cu3 distances remain nearly the same, and upon further
cooling to 130 K, they are shortened by about 0.007 ꢁ
(Supporting Information, Figure S1). It is noteworthy that
the Cu3(m2-Cl)3 core was previously found only in two
complexes shown in Figure 1.[22] Interestingly, a related Cu3
complex bearing the [o-(Ph2P)C6H4]3P ligand does not exhibit
m3-bridging coordination of the central P atom (Cu–P 2.889,
2.889 and 2.723 ꢁ, av. Cu···Cu 2.52 ꢁ),[26] possibly owing to
steric reasons.
Results and Discussion
It has been previously reported[24] that the phosphine L is
readily assembled with 2 equiv of CuI or AgI halide into 1D
coordination polymers of [M4Hal4L2]n formula (M = Cu or
Ag, and Hal = Cl, Br or I). Reproducing these syntheses, we
Figure 2. Molecular structures of 1–3 determined by sc-XRD analysis.
Angew. Chem. Int. Ed. 2021, 60, 2 – 10
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
&&&&
These are not the final page numbers!

Angewandte
Research Articles
Chemie
The compounds 1–3 are colorless crystals or white
powders, moderately soluble in acetonitrile. Bromide 2 and
especially iodide 3 tolerate air atmosphere well, while
chloride 1 is slowly oxidized under ambient conditions. The
phase identities of the synthesized samples were verified by
powder XRD (Supporting Information, Figures S3–S5),
FTIR spectroscopy, and microanalysis data. Surprisingly,
unlike most neutral CuI halide complexes, compounds 1–3
demonstrate enhanced thermal stability remaining pristine up
to about 2708C (Supporting Information, Figure S6).
The solution behavior of complexes 1–3 was studied by
NMR spectroscopy (Supporting Information, Figures S10–
12). The 1H NMR spectra reveal a single set of ligand
multiplets, which are significantly downfield-shifted in com-
parison with those of the free ligand L. Interestingly, the
methylene protons of CH2Py groups of 1–3 appear as
a doublet of triplets, while in free L, they resonate as
a multiplet. The 31P{1H} spectra of the complexes show
a singlet peaking at the range from ꢀ42.70 (1) to ꢀ65.74 (3)
[cf. dP = ꢀ28 ppm for free ligand]. Taken together, these data
imply that the above complexes 1–3 do not dissociate upon
dissolution, which is consistent with the computational results
(see below).
The molecular geometries of complexes 1–3, optimized in
the gas phase at the B97-D3 level, are in good agreement with
those obtained from sc-XRD analysis (Table 1; Supporting
Information, Tables S2, S3), and the calculated IR spectra
coincide well with the experimental ones (Supporting In-
formation, Figure S13). The results of calculations for iodide
complex 3 are representative. For this complex, the difference
between the experimental structure and the optimized one is
most noticeable, namely, the deviation of the optimized
structure from the C3 symmetry is insignificant (Table 1).
Electronic structure and bonding interactions have been
analyzed for the sc-XRD geometries of the complexes using
the QTAIM,[28] NBO[30] and NEDA[31] procedures. Table 1
shows that the Cu–I, Cu–N, and Cu–P coordination bonds in 3
are characterized by a low electron density at the bond critical
points (BCPs), which are in the range of 0.03–0.09 a.u. For all
these BCPs, the electron density Laplacian is positive. It
differs significantly for the bonds of different types, while for
the bonds of the same type it correlates with their length and
bond order (Table 1). The values of the discussed topological
descriptors are typical of closed-shell (CS) bonding interac-
tions, although the j Vb j /Gb values are slightly higher than
unity, thus indicating their partial covalence.[32]
The most interesting is the first discovered m3-P-bridging
coordination with significantly different Cu–P distances.
Nevertheless, BCPs are localized even for the longest Cu–P
distances (e.g. 2.84 ꢁ in 3). This fact gives evidence of the
formation of a coordination bond, although with a very low
bond order (0.13 in 3). The sum of the Myers bond orders for
three Cu-P coordination bonds is close to unity (ca. 0.84). The
Cu···Cu interactions also exhibit a low bond order (Pav ꢁ 0.15).
Other coordination bonds, namely, Cu–N and Cu–I, have
much higher bond orders (ca. 0.5–0.6). Note that the Mayer
bond order for typical P-C covalent bonds is close to unity
(Table 1).
According to the natural energy decomposition analysis
(NEDA), the energy of interaction between the CuI cations,
halide ions, and the ligand L is 647 kcalmolꢀ1 for complex 3
(663 and 682 kcalmolꢀ1 for 1 and 2, respectively). The charge
transfer (or orbital) contribution (ECT) to this value for 3 is
624.2 kcalmolꢀ1, which is 96% of the total bonding energy.
Similar results were obtained for complexes 1 and 2 (for
details, see the Supporting Information, Section 7.3). Thus, to
estimate the energy of all coordination bonds (Table 1;
Supporting Information, Tables S2, S3), we performed NBO
analysis, which treated each complex (1–3) as seven inde-
pendent units (namely, three CuI cations, three halides, and
the ligand L). The charge transfer contributions (ECT) to the
bond energies were calculated using the second-order pertur-
bation theory (Table 1; Supporting Information, Tables S2
and S3).
Figure 3 displays the pairs of the Lewis donor and non-
Lewis acceptor NBOs, the interactions of which contribute
mainly to the energies of the considered coordination bonds.
Note that the same lone pair (LP) on the phosphorus atom
interacts with the valence 4s-orbital of each Cu ions. Taken
together, these interactions may be viewed as a four-center
two-electron interaction, resulting from the donation of
a phosphorus lone pair into three vacant copper 4s orbitals.
It should be clarified that a smaller, albeit noticeable,
contribution (ca. 10%) to the P-Cu bonding is made by the
interactions of the same LP with the Rydberg orbitals of Cu.
In addition, the PCu3 core is also stabilized by a series of Cu!
Table 1: Selected experimental and calculated[a] bond lengths (rb, ꢀ) in complex 3, sums of the covalent radii[27] of the bonded atoms (rcov, ꢀ), the
QTAIM topological descriptors:[28] electron density (1b), its Laplacian (D1b), and ratio of the potential Vb and kinetic Gb energy densities (jVb j/Gb) at
the bond critical points (BCPs),[b] Mayer bond orders (P) and orbital (or CT) contributions (ECT) to the bond energies, estimated using the NBO
analysis and second-order perturbation theory.
P
P
Eð2Þ[c]
2
Bond
rb
rcov
1b
r 1b
jVb j/Gb
P
ECT
=
XRD
B97-D3
Cu-I
2.59 to 2.68
2.68
2.58
2.05
2.55
1.88
2.71
2.39
2.03
2.64
1.83
0.053 to 0.045
0.059 to 0.030
0.085 to 0.083
0.040 to 0.033
0.162 to 0.158
0.116 to 0.045
0.12 to 0.06
0.38 to 0.36
0.099 to 0.069
ꢀ0.28 to ꢀ0.25
1.13 to 1.11
1.32 to 1.17
ca. 1.20
1.19 to 1.21
2.91 to 2.70
0.59 to 0.47
0.41, 0.30, 0.13
0.53 to 0.49
0.17 to 0.13
0.98 to 0.88
60.9 to 51.8
34.6, 23.8, 10.7
39.2, 36.4, 30.6
2.6 to 2.9
Cu-P
Cu-N
Cu···Cu
P-C
2.41, 2.53, 2.84
2.01 to 2.05
2.47 to 2.59
1.85 to 1.87
[a] Geometry optimization was performed at the B97-D3/def2-TZVP (with ECP for I) level.[29] [b] Based on the gas-phase B3LYP/def2-TZVP (with an
ECP for I) calculations at XRD geometry; values of 1 and D1 are given in a.u., that is, e/a 3 and e/a0 , respectively. [c] Every contribution is evaluated
5
ꢀ ꢀ
ꢁ
ꢂ
b
b
0
2
ꢀ ꢀ
b
using the second-order perturbation theory as Eð2Þ
¼
nDhFD F F i =ðeD ꢀ eAÞ, where FD and FA, and eD and eD are the NBOs of donor and
ꢀ ꢀ
A
acceptor units and their energies, respectively;[30] nD is the occupation number of the donor NBO, and F is a Fock operator.
b
&&&&
www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 2 – 10
These are not the final page numbers!

Angewandte
Research Articles
Chemie
Figure 3. Representative pairs of strongly interacting NBOs [isosurface at 0.08 a.u.; the donor Lewis NBOs (D): lone pairs (LP) of I, P and N
atoms and 3d valence orbital of Cu; the acceptor non-Lewis NBOs (A): 4s orbital of Cu] and the energies of their interactions calculated using
second order perturbation theory.
P interactions with the participation of the 3d-orbitals of Cu
ions as donors and vacant s*(P-C) as acceptors (ca. 12%).
The formation of a four-center two-electron bonding is also
confirmed by NLMO analysis[30] (Supporting Information,
Table S5, Figure S16).
The energies of considered coordination bonds are
calculated as the sum of all orbital contributions and are
presented in Table 1 and the Supporting Information, Ta-
bles S2 and S3 for 1–3, respectively. As seen from the Tables,
the Cu–X coordination bonds are characterized by the highest
energy values. The nature of the halogen has little effect, and
the change in these values in the range of 50–61 kcalmolꢀ1
mainly composed of 5p orbitals of iodine atoms with a smaller
contribution of 3d-AOs of Cu atoms. In turn, the six lowest-
energy unoccupied MOs (LUMO to LUMO + 5) are local-
ized on the pyridine (p*-MOs of pyridines). The calculations
of electronic absorption spectra of 1–3 (Supporting Informa-
tion, Section 7.4 and Figures S17, S18) at the TD-CAM-
B3LYP level predict that a series of electronic transitions in
the 250–360 nm range is of the (M + X)LCT (metal-to-ligand
and halide-to-ligand) type.
To shed light on the luminescence properties of 1–3 (vide
infra), we performed the geometry optimization for 1–3 in the
triplet state. At the B97-D3 level, we were able to optimize
3
ꢀ
depends mainly on the Cu X bond length for the same X. The
only geometries of 1–3 in the (M + X)LCT state (T1). The
ꢀ
Cu N bonds have the next largest orbital (or CT) contribu-
triplet-singlet energy gaps DE(T1ꢀS0) calculated at the CAM-
B3LYP level correspond to the phosphorescence maxima
equal to 431–443 nm, which is in excellent agreement with the
experimental data for 1–3 (Table 2).
tion, which varies from 46.3 to 30.6 kcalmolꢀ1 for the bond
ꢀ
lengths from 1.99 to 2.05 ꢁ. As mentioned above, the Cu P
bond lengths in complexes 1 and 3 differ greatly (from 2.41 to
2.90 ꢁ), which significantly changes their energies (34.6–
7.3 kcalmolꢀ1). The orbital contribution to the energy of the
Cu···Cu short contacts is very low (3.7–2.6 kcalmolꢀ1).
At the B3LYP level, we succeeded in optimizing the
geometries of the two types of triplet states for each complex:
T1 of the 3[(M + X)LCT] type and T2. The latter triplets
correspond to the electron excitation from HOMO, consisting
of p-AOs of halogens and 3d-AOs of Cu, to a molecular
orbital, composed mainly of 4s/4p orbitals of Cu and P atoms
(Supporting Information, Figures S19, S20). Thus, these T2
Moreover, to understand the nature and origin of the
electronic transitions responsible for absorption and lumi-
nescence spectra of 1–3, their ground and lowest excited
states were considered. DFT calculations predict that the nine
highest occupied MOs (HOMO to HOMOꢀ8) of 3 are
3
triplets are of the CC type.[33,34] The band maxima for both
Table 2: Emission maxima (lem), PL lifetimes, and PL quantum yields (PLQYs) at 300 and 77 K as well as calculated (M+X)LCT type phosphorescence
maxima (lcpahlc) for 1–3.
calc
ph
Complex
lem [nm]
Lifetime [ms]
77 K
PLQY, %
[300 K]
l
[nm][d]
(HE emission)
300 K
475
77 K
300 K
15.0
[Cu3(m2-Cl)3L] (1)[a]
(powder or crystals)
[Cu3(m2-Br)3L] (2)[a]
(powder or crystals)
[Cu3(m2-I)3L] (3)[b]
(pristine crystals)
Ground 3 (3gr)[b]
495 (HE)
580 (LE)[c]
461 (HE)
580 (LE)[c]
462 (HE)
622 (LE)
450 (HE)
505 (HE1)
610 (LE)
505 (HE)
43.3 (HE)
20 (LE)
56
443
431
435
-
453
13.4
65.8 (HE)
13 (LE)[c]
48.0 (HE)
10 (LE)
100
50
442 (HE)
613 (LE)
505
11.0 (HE)
5 (LE)
16.5
50 (HE)
22
128 (HE1)
11 (LE)
Melted 3 (3mt)[b]
(amorphous)
505 (HE1)
614 sh (LE)
19.8 (HE1)
5.5 (LE)
170 (HE1)
30
-
[a] The data given are recorded at lex =370 nm; [b] The data given are recorded at lex =360 nm; [c] The LE band is observed at lex >420 nm;
[d] Calculated at the (U)CAM-UB3LYP/def2-TZVP level (for details, see Section 7.5 in the Supporting Information).
Angew. Chem. Int. Ed. 2021, 60, 2 – 10
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
&&&&
These are not the final page numbers!

Angewandte
Research Articles
Chemie
types of phosphorescence were estimated (Supporting In-
formation, Table S7) and were in satisfactory agreement with
experiment, although the CC-type phosphorescence is sig-
nificantly red-shifted. According to the calculations, the T1
state is lower than T2 for all complexes, but the energy
difference decreases significantly (by about 870 cmꢀ1) from
1 to 3 (see the Supporting Information, Section 7.5). This
trend explains the difference in thermochromic properties of
1–3 (see below).
Upon UV irradiation at 300 K, the pristine crystals of 1–3
emit a bright sky-blue, deep-blue, and a bluish–white
luminescence, respectively. It should be highlighted that the
deep-blue and near-white luminescence, especially highly
efficient one, are extremely rare observed among CuI com-
plexes. It is thus noteworthy that the photoluminescence
quantum yield (PLQY) of deep-blue emitting complex 2
(lmax = 453 nm) is 100% at 300 K, with a lifetime being
13.4 ms. The complexes 1 and 3 also demonstrate a high
quantum performance of 56 and 50%, respectively. The
temperature-dependent emission spectra of 1–3 measured in
the 77–300 K window are shown in Figure 4, and the
corresponding excitation profiles are plotted in the Support-
ing Information, Figures S21, S23, S25. The photophysical
data for 300 and 77 K are outlined in Table 2. As seen from
Figures 4a,b, the emission spectra of 1 and 2 recorded at
300 K (lex = 360 nm) reveal a broad band with maxima at 475
and 453 nm, respectively. Upon cooling from 300 to 77 K,
emission profiles of 1 and 2 are gradually red-shifted by 8–
20 nm, which slightly changes the emission chromaticity
(Figure 4d). In contrast, the iodide 3 at the same excitation
wavelength features remarkable dual emission expressed by
high-energy (HE) and low-energy (LE) bands, whose relative
ratio is temperature-sensitive (Figure 4c). At 300 K and lex =
360 nm, the luminescence of 3 is largely contributed by the
HE band (lmax = 442 nm) with a small admixture of the LE
component (lmax = 613 nm). With decreasing temperature up
to ꢁ 160 K, the intensity of LE component significantly
increases with a concomitant decrease of HE component.
With further temperature decay, the LE component starts to
grow again (Figure 4c). Thus, the integral intensity ratio of
HE and LE components (IHE/ILE, Figure 4 f) dramatically
grows from 300 to about 160 K, and, upon further cooling to
77 K, it returns to nearly the same magnitude as that at 240 K.
As a result, complex 3 exhibits non-trivial thermochromic
luminescence illustrated on the CIE diagram (Figure 4d), and
visualized in the Supporting Information, Movie). Namely,
upon cooling from 300 to about 155 K, the emission color of 3
shifts from a bluish-white to an orange, and upon further
cooling to 77 K, it becomes a bluish-white again. Note that
this phenomenon is completely reversible: according to the
temperature-dependent sc-XRD study (see above), the com-
plex 3 does not undergo any structural transitions over the
entire temperature range. To our knowledge, the observed
back-and-forth pattern of the chromaticity changing, associ-
3
Figure 4. a)–c) Temperature-dependent emission spectra of pristine 1 (a), 2 (b), and 3 (c), recorded at lex =360 nm; d) Changing of the emission
chromaticity of pristine 1–3 upon cooling down from 300 to 77 K (lex =360 nm); e) Excitation spectra for pristine 1–3 at 77 K; f) Temperature
dependence of the HE to LE emission band integral ratio (ILE/IHE) of pristine 3 at lex =360 nm.
&&&&
www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 2 – 10
These are not the final page numbers!

Angewandte
Research Articles
Chemie
ated with the thermochromic luminescence, is unprecedented
grinding (for details, see the Supporting Information, Sec-
for Group 11 metal complexes. Together with this effect,
complex 3 displays an excitation-wavelength dependent
(EWD) luminescence that appears as changing the HE/LE
band ratio, and hence emission color, upon variation of
excitation energy (Supporting Information, Figures S26, S27).
Therefore, the emission color of 3 can be precisely tuned
across the visible range through the altering of both temper-
ature and excitation energy.
The dual emission property of 3 clearly suggests two
different excited states, which are responsible for the HE and
LE emission bands. The same is true for complexes 1 and 2,
whose weak LE emission band (lmax = 580 and 622 nm) begins
to appear at low energy excitations (> 410 nm). Unlike 3, the
LE bands of both 1 and 2 are much less pronounced as
compared with the dominating HE band (Supporting Infor-
mation, Figures S22–S24). Again, the emission color of 1 and
2 can be finely adjusted in a wide visible range through the
altering of the excitation wavelength (Supporting Informa-
tion, Figures S22 and S24).
tion 9). The PXRD, FT-IR and emission patterns of the
recovered phase 3 coincide with those for the pristine one,
indicating the chemical identity of both.
As seen in Figure 5b, grinding of 3 significantly perturbs
its emission profile at ambient temperature, causing a new
dominant band at lmax = 505 nm (henceforth referred to as
HE1), which is consistent with the visible appearance of green
emission. However, in the emission spectra of 3gr recorded at
lower temperatures (Figure 5b; Supporting Information, Fig-
ure S31a), the HE1 band is flanked by HE and LE bands,
whose maxima, lifetimes (Table 2) and excitation profiles are
about the same as those of pristine 3. Given this finding, one
can conclude that powder 3gr contains the pristine (crystal-
line) 3 phase. Moreover, the low-temperature emission data
indicates that a more thoroughly grinding of 3gr does not lead
to the complete disappearance of the crystal phase. Since 3
melts without decomposition at about 1818C (Supporting
Information, S6, S9), forming an amorphous phase (3mt), we
assumed that the latter should display only the emission bands
characteristic of a fully amorphized complex 3 (for details, see
the Supporting Information, Section 9). Indeed, the emission
spectra of 3mt recorded at 300 K represent overlapping bands
of the dominant HE1 (lmax = 505 nm) and LE (lmax ꢁ 614 nm)
components (Figure 5c; Supporting Information, Figure S33),
the lifetimes of which are 19.8 and 5.5 ms, respectively
(Table 2). An increase in temperature to 350 K equalizes
these bands, while cooling to 77 K completely suppresses the
LE band (Figure 5c).
The luminescence decay of 1–3 in the maxima of HE and
LE bands has an monoexponential character throughout 77–
300 K range. At 300 K, the lifetimes of the HE emission (tHE
)
drop in the order 1 (15.0 ms) > 2 (13.4 ms) > 3 (11.0 ms), which
correlates with the heavy atom effect. Upon lowering the
temperature to 77 K, tHE rises up to 43.3–50 ms (Table 2). The
shape of the tHE(T) dependences (Supporting Information,
Figure S28) precludes the TADF (thermally activated de-
layed fluorescence) character of the HE emission, thereby
clearly indicating its phosphor-
escence origin. The lifetimes of
the LE emission of 1–3 being
5–20 ms (Table 2) also implies
the phosphorescence nature of
this component.
Furthermore, complex
3
features intriguing mechano-
chromic luminescence, while
its analogs 1 and 2 do not show
this rare property. A soft pes-
tling of pristine 3 in a mortar
cause drastic changes in the
emission color from bluish–
white to green (Figure 5a, i–
iii). The treatment of the
crushed powder (3gr) with
a few drops of acetonitrile fully
recovers the original emission
of the pristine phase 3 (Fig-
ure 5a, iv,v). Again, a pestle
leaves greenish emission traces
on the surface of the recovered
powder 3 (Figure 5a, vi). The
PXRD and FTIR analyses of
the pristine (3) and ground
(3gr) phase yield nearly the
same patterns, thus giving evi-
dence that no chemical trans-
formations take place upon
Figure 5. Mechanochromic luminescence of 3: a) Images obtained under UV-lamp: i)–iii) Changing the
emission color of the intact microcrystals 3 upon grinding with a pestle; iv),v) a reversion to the bluish-
white emission by drop-wise treatment of 3gr (ground 3) with MeCN; vi) scratching the powder 3 formed
with a pestle. b) Emission spectra of 3, 3gr, and 3mt (melted 3) at 300 and 77 K (lex =360 nm);
c) Emission spectra of 3mt recorded in the 77–350 K window (lex =360 nm).
Angew. Chem. Int. Ed. 2021, 60, 2 – 10
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
&&&&
These are not the final page numbers!

Angewandte
Research Articles
Chemie
possible only above about
180 K. The amorphization-
forced shifting of the 3(M +
X)LCT potential energy sur-
face also increases the Eaback
barrier so that the reverse 3CC
!
3(M + X)LCT population
does not occur even at 350 K.
The assignment of the HE1
3
band of 3gr and 3mt to (M +
X)LCT one is corroborated by
the similarity of its excitation
spectra (Supporting Informa-
tion, Figures S31c, S33e) and
lifetimes (Table 2) with those
Figure 6. Simplified diagrams explaining the thermo- and mechanochromic luminescence of 3. Left:
pristine 3, right: fully amorphized 3, that is, 3mt.
Comparing the above photophysical data with the results
of our DFT calculations and literature data for polynuclear
CuI-based clusters,[33–38] it is natural to assign the HE bands of
1–3 to phosphorescence from the 3(M + X)LCT state, and the
LE bands to the 3CC phosphorescence. Moreover, the
of the HE band of pristine 3 (Supporting Information,
Figure S25a).
Therefore, our findings give evidence that the m3-P-
coordinated ligand-derived complex 3 is endowed with an
amazing combination of the following properties: a) non-
trivial thermochromic luminescence (an unprecedented back-
and-forth changing chromaticity pattern, Figure 4d); b) out-
standing mechanochromic luminescence (Dlmax ꢁ 60 nm upon
soft grinding, Figure 5b); and c) pronounced excitation-
dependent behavior.
To our knowledge, such pronounced mechanochomic
luminescence have never been described for CuI-based
compounds. The previous record belongs to the cluster
[Cu4I4(4-dpda)4] [4-dpda = 4-(diphenylphosphino)-N,N-dime-
3
lifetimes of CC states are typically much shorter than those
3
of (M + X)LCT ones,[33–38] and the character of the tHE(T)
dependences rules out the TADF manifestation.[39]
To explain the remarkable dual luminescence of 3, we
propose the following simplified energy diagram (Figure 6).
As the UV spectra of 1–3 are due to transitions to the 1(M +
X)LCT states, the 3CC state must be populated from the
3(M + X)LCT one through the thermally activated chan-
nel.[33–38] The energy barrier of this process (Edair), estimated
from the Arrhenius plot for the IHE/ILE ratio is about 220 cmꢀ1
(Supporting Information, Figure S29 and Section 8.1). Owing
thylaniline], demonstrating an emission shift of 50 nm (Dlmax
)
upon grinding.[40]
3
to the small Edair barrier, the CC state is populated from the
3(M + X)LCT level already at 77 K, so that both LE and HE
emission bands are observed (Figure 4c). When the temper- Conclusion
ature rises from 77 to about 160 K, a higher thermal
3
population of the CC state occurs that results in increasing
of the ILE/IHE ratio (Figure 4 f). However, further warming
leads to reversibility of conversion between 3(M + X)LCTand
3CC states. Consequently, in the 160–300 K window, the HE
band increases. Most likely, the second linear segment on the
Arrhenius plot (Supporting Information, Figure S29) gives
A m3-P-bridging coordination pattern of a phosphine
ligand has been discovered for the first time. Namely, in
a series of [Cu3(m2-Hal)3L] (Hal = Cl, Br, or I) complexes,
supported by tris[2-(2-pyridyl)ethyl]phosphine (L), the phos-
phorus atom bridges three Cu atoms of the Cu3 triangle. The
structural evidence of the m3-P-bridging pattern disclosed is
strongly corroborated by theoretical calculations, which
justify a four-centre two-electron bonding interaction of the
phosphorus atom with the three CuI atoms. Moreover, the
compounds studied display blue or bluish-white phosphor-
escence (lem = 442–465 nm) with the quantum efficiency up to
50–100%. A remarkable and fully unprecedented thermo-
and mechanochromic luminescence is also discovered for
[Cu3(m2-I)3L] complex. Namely, changing its emission chro-
maticity upon temperature varying within 77–300 K window
looks like a back-and-forth pattern on the CIE diagram. Upon
a soft pestling of this complex, its emission color sharply
changes from bluish–white to green and recovers again upon
treatment of the ground powder with acetonitrile. Thus, the
emission color of the iodide complex can be fine-tuned in
a wide chromaticity window through three different tools:
temperature, mechanical stress and excitation energy.
3
3
the energy difference between (M + X)LC and CC states
(Eaback ꢀ Eadir), estimated as 840 cmꢀ1. As discussed above,
DFT calculations predict for 3 the negative sign of this
difference (ꢀ667 cmꢀ1; Supporting Information, Table S8)
due to the typical underestimation of the energy of the CT
states. Assuming the same extent of this underestimation for
all complexes, we can infer that (Ebaack ꢀ Eadir) value decreases
significantly for 2 and vanishes for 1. This decrease, in turn,
should lead to a significant increase in the value of Edair, which
may explain the absence of LE luminescence for 1 and 2 (for
details, see the Supporting Information, Section 7.5).
Grinding (3!3gr) or melting (3/3gr!3mt) leads to
a red-shift of the HE emission profile (Figure 5b,c), thus
a minimum of the 3(M + X)LCT potential shifts closer to the
S0 surface, causing increase of the Edair barrier from 220 to
570 cmꢀ1 (Supporting Information, Figure S33b). Thereupon,
the thermal population of the 3CC state from the 3(M +
X)LCT level is hampered at low temperature and becomes
Our findings also contribute significantly to the funda-
mental chemistry of pnictine(III)-based ligands, chemical
&&&&
www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 2 – 10
These are not the final page numbers!

Angewandte
Research Articles
Chemie
bond theory, and photophysics of CuI compounds. From an
application viewpoint, the described compounds may be of
interest both as promising stimuli-responsive materials and
potential emitters for OLED applications.[41]
[17] S. Welsch, C. Lescop, G. Balazs, R. Rꢂau, M. Scheer, Chem. Eur.
J. 2011, 17, 9130 – 9141.
[18] C. Lescop in Copper(I) Chemistry of Phosphines, Functionalized
Phosphines and Phosphorus Heterocycles (Ed.: M. S. Balakrish-
na), Elsevier, Amsterdam, 2019, pp. 21 – 59.
[19] Y. Hou, Z. Li, Y. Li, P. Liu, C.-Y. Su, F. Puschmann, H.
Grꢃtzmacher, Chem. Sci. 2019, 10, 3168 – 3180.
Acknowledgements
[20] A. L. Balch, B. J. Davis, M. M. Olmstead, J. Am. Chem. Soc.
1990, 112, 8592 – 8593.
[21] A. L. Balch, B. J. Davis, M. M. Olmstead, Inorg. Chem. 1993, 32,
3937 – 3942.
This work was supported by Russian Science Foundation
(Project 18-73-10086). E.A.P. and N.P.G. also acknowledge
Russian Science Foundation (Project 16-13-10155) for a fi-
nancial support of the computational part of this work and the
Supercomputer Centre of Novosibirsk State University for
computational resources.
[22] I. S. Ke, F. P. Gabbaꢀ, Aust. J. Chem. 2013, 66, 1281 – 1287.
[23] S. L. Benjamin, T. Krꢄmer, W. Levason, M. E. Light, S. A.
Macgregor, G. Reid, J. Am. Chem. Soc. 2016, 138, 6964 – 6967.
[24] A. V. Artemꢅev, A. Y. Baranov, M. I. Rakhmanova, S. F. Maly-
sheva, D. G. Samsonenko, New J. Chem. 2020, 44, 6916 – 6922.
[25] A. Bondi, J. Phys. Chem. 1964, 68, 441 – 451.
[26] A. Belyaev, T. M. Dau, J. Jꢄnis, E. V. Grachova, S. P. Tunik, I. O.
Koshevoy, Organometallics 2016, 35, 3763 – 3774.
[27] B. Cordero, V. Gꢆmez, A. E. Platero-Prats, M. Revꢂs, J.
Echeverrꢇa, E. Cremades, F. Barragꢈn, S. Alvarez, Dalton Trans.
2008, 2832 – 2838.
Conflict of interest
The authors declare no conflict of interest.
[28] a) C. F. Matta, R. J. Boyd, The Quantum Theory of Atoms in
Molecules, Wiley, Weinheim, 2007; b) R. F. W. Bader, Monatsh.
Chem. 2005, 136, 819 – 854; c) R. F. W. Bader, Atoms in Mole-
cules: A Quantum Theory, Oxford University Press, Oxford,
1990.
Keywords: coordination modes · metal–metal interactions ·
multiple bonds · phosphane ligands · photophysics
[29] Detailed information on the employed quantum chemical
methods and programs can be found in the Supporting
Information (Section 7.1).
[1] F. A. Cotton, G. Wilkinson, C. A. Murillo, M. Bochmann,
Advanced Inorganic Chemistry, 6nd ed., Wiley, New York,
1999, chap. 16.4.
[30] a) A. E. Reed, L. A. Curtiss, F. Weinhold, Chem. Rev. 1988, 88,
899 – 926; b) F. Weinhold, C. R. Landis, E. D. Glendening, Int.
Rev. Phys. Chem. 2016, 35, 399 – 440.
[31] E. D. Glendening, C. R. Landis, F. Weinhold, J. Comput. Chem.
2013, 34, 1429 – 1437.
[2] Comprehensive Coordination Chemistry II (Eds.: J. A. McClev-
erty, T. J. Meyer), Elsevier, Oxford, 2004.
[3] Phosphorus(III) Ligands in Homogeneous Catalysis: Design and
Synthesis (Eds.: P. C. J. Kamer, P. W. N. M. van Leeuwen), Wiley,
Hoboken, 2012.
[32] a) E. Espinosa, I. Alkorta, J. Elguero, E. Molins, J. Chem. Phys.
2002, 117, 5529 – 5542; b) W. Nakanishi, S. Hayashi, K. Narahara,
J. Phys. Chem. A 2009, 113, 10050 – 10057.
[33] F. De Angelis, S. Fantacci, A. Sgamellotti, E. Cariati, R. Ugo,
P. C. Ford, Inorg. Chem. 2006, 45, 10576 – 10584.
[34] S. Perruchas, C. Tard, X. F. Le Goff, A. Fargues, A. Garcia, S.
Kahlal, J. Y. Saillard, T. Gacoin, J. P. Boilot, Inorg. Chem. 2011,
50, 10682 – 10692.
[35] B. Huitorel, R. Utrera-Melero, F. Massuyeau, J. Y. Mevelec, B.
Baptiste, A. Polian, T. Gacoin, C. Martineau-Corcos, S. Perru-
chas, Dalton Trans. 2019, 48, 7899 – 7909.
[4] P. Schwab, N. Mahr, J. Wolf, H. Werner, Angew. Chem. Int. Ed.
Engl. 1994, 33, 97 – 99; Angew. Chem. 1994, 106, 82 – 84.
[5] T. Pechmann, C. D. Brandt, H. Werner, Angew. Chem. Int. Ed.
2000, 39, 3909 – 3911; Angew. Chem. 2000, 112, 4069 – 4072.
[6] T. Pechmann, C. D. Brandt, C. R. Rçger, H. Werner, Angew.
Chem. Int. Ed. 2002, 41, 2301 – 2303; Angew. Chem. 2002, 114,
2398 – 2401.
[7] P. Schwab, J. Wolf, N. Mahr, P. Steinert, U. Herber, H. Werner,
Chem. Eur. J. 2000, 6, 4471 – 4478.
[8] T. Pechmann, C. D. Brandt, H. Werner, Chem. Commun. 2003, 3,
1136 – 1137.
[36] R. Utrera-Melero, B. Huitorel, M. Cordier, J. Y. Mevellec, F.
Massuyeau, C. Latouche, C. Martineau-Corcos, S. Perruchas,
Inorg. Chem. 2020, 59, 13607 – 13620.
[37] A. Kobayashi, Y. Yoshida, M. Yoshida, M. Kato, Chem. Eur. J.
2018, 24, 14750 – 14759.
[38] P. C. Ford, E. Cariati, J. Bourassa, Chem. Rev. 1999, 99, 3625 –
3647.
[39] R. Czerwieniec, M. J. Leitl, H. H. H. Homeier, H. Yersin, Coord.
Chem. Rev. 2016, 325, 2 – 28.
[9] T. Pechmann, C. D. Brandt, H. Werner, Dalton Trans. 2004, 6,
959 – 966.
[10] T. Pechmann, C. D. Brandt, H. Werner, Chem. Eur. J. 2004, 10,
728 – 736.
[11] H. Werner, Angew. Chem. Int. Ed. 2004, 43, 938 – 954; Angew.
Chem. 2004, 116, 956 – 972.
[12] M. Sauthier, B. Le Guennic, V. Deborde, L. Toupet, J.-F. Halet,
R. Rꢂau, Angew. Chem. Int. Ed. 2001, 40, 228 – 231; Angew.
Chem. 2001, 113, 234 – 237.
[40] X.-C. Shan, F.-L. Jiang, L. Chen, M.-Y. Wu, J. Pan, X.-Y. Wan,
M.-C. Hong, J. Mater. Chem. C 2013, 1, 4339 – 4349.
[41] W. Liu, Y. Fang, J. Li, Adv. Funct. Mater. 2018, 28, 1705593.
[13] F. Leca, C. Lescop, E. Rodriguez-Sanz, K. Costuas, J. F. Halet, R.
Rꢂau, Angew. Chem. Int. Ed. 2005, 44, 4362 – 4365; Angew.
Chem. 2005, 117, 4436 – 4439.
[14] B. Nohra, E. Rodriguez-Sanz, C. Lescop, R. Rꢂau, Chem. Eur. J.
2008, 14, 3391 – 3403.
[15] S. Welsch, C. Lescop, M. Scheer, R. Rꢂau, Inorg. Chem. 2008, 47,
8592 – 8594.
Manuscript received: March 1, 2021
[16] S. Welsch, B. Nohra, E. V. Peresypkina, C. Lescop, M. Seheer, R.
Rꢂau, Chem. Eur. J. 2009, 15, 4685 – 4703.
Accepted manuscript online: March 22, 2021
Version of record online: && &&
,
&&&&
Angew. Chem. Int. Ed. 2021, 60, 2 – 10
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
&&&&
These are not the final page numbers!

Angewandte
Research Articles
Chemie
Research Articles
Coordination Modes
A. Y. Baranov, E. A. Pritchina,
A. S. Berezin, D. G. Samsonenko,
V. P. Fedin, N. A. Belogorlova,
N. P. Gritsan,*
A. V. Artem’ev*
&&&& — &&&&
Beyond Classical Coordination
Chemistry: The First Case of a Triply
Bridging Phosphine Ligand
Phosphines are notorious in acting
ligand bridges three metal atoms at once.
Such triply bridging pattern is first
observed in the [Cu3(m2-Hal)3L] (Hal=Cl,
Br, or I) complexes, which also exhibit
remarkable luminescence properties.
exclusively as terminal donors, binding
a single metal center. We have discovered
that a non-traditional scenario is possible
where the phosphorus(III) atom of a R3P
&&&&
www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 2 – 10
These are not the final page numbers!