Understanding the Excited State Behavior of Cyclometalated Bis(tridentate)ruthenium(II) Complexes: A Combined Experimental and Theoretical Study

Article abstract of DOI:10.1021/acs.inorgchem.5b01151

The synthesis and characterization of the donor-acceptor substituted cyclometalated ruthenium(II) polypyridine complex isomers [Ru(dpb-NHCOMe)(tpy-COOEt)](PF₆) 1(PF₆) and [Ru(dpb-COOEt)(tpy-NHCOMe)](PF₆) 2(PF₆)

Full text of DOI:10.1021/acs.inorgchem.5b01151

This is an open access article published under an ACS AuthorChoice License, which permits
copying and redistribution of the article or any adaptations for non-commercial purposes.
Article
pubs.acs.org/IC
Understanding the Excited State Behavior of Cyclometalated
Bis(tridentate)ruthenium(II) Complexes: A Combined Experimental
and Theoretical Study
Christoph Kreitner,† Elisa Erdmann, Wolfram W. Seidel, and Katja Heinze*
,‡
§
§
,†
†Institute of Inorganic and Analytical Chemistry, Johannes Gutenberg-University of Mainz, Duesbergweg 10-14, 55128 Mainz,
Germany
‡
Graduate School Materials Science in Mainz, Staudingerweg 9, 55128 Mainz, Germany
§
Institute of Chemistry, University of Rostock, Albert-Einstein-Straße 3a, 18059 Rostock, Germany
*
S Supporting Information
ABSTRACT: The synthesis and characterization of the
donor−acceptor substituted cyclometalated ruthenium(II)
polypyridine complex isomers [Ru(dpb-NHCOMe)(tpy-
COOEt)](PF ) 1(PF ) and [Ru(dpb-COOEt)(tpy-
6
6
NHCOMe)](PF ) 2(PF ) (dpbH = 1,3-dipyridin-2-ylbenzene,
6
6
tpy = 2,2′;6,2″-terpyridine) with inverted functional group
pattern are described. A combination of resonance Raman
spectroscopic and computational techniques shows that all
intense visible range absorption bands arise from mixed Ru →
tpy/Ru → dpb metal-to-ligand charge transfer (MLCT)
excitations. 2(PF ) is weakly phosphorescent at room
6
temperature in fluid solution and strongly emissive at 77 K
in solid butyronitrile matrix, which is typical for ruthenium(II)
polypyridine complexes. Density functional theory calculations revealed that the weak emission of 2(PF ) arises from a MLCT
state that is efficiently thermally depopulated via metal-centered ( MC) excited states. The activation barrier for the deactivation
process was estimated experimentally from variable-temperature emission spectroscopic measurements as 11 kJ mol . In
3
6
3
−1
contrast, 1(PF ) is nonemissive at room temperature in fluid solution and at 77 K in solid butyronitrile matrix. Examination of
6
3
the electronic excited states of 1(PF ) revealed a ligand-to-ligand charge-transfer ( LL′CT) as lowest-energy triplet state due to
6
the very strong push−pull effect across the metal center. Because of the orthogonality of the participating ligands, emission from
3
the LL′CT is symmetry-forbidden. Hence, in this type of complex a stronger push−pull effect does not increase the
phosphorescence quantum yields but completely quenches the emission.
INTRODUCTION
has a long excited-state lifetime of 855 ns (in acetonitrile) due
to the spin-forbidden character of the luminescence.
■
5
Polypyridine complexes of ruthenium have been studied
extensively in the last 50 years. Especially, the photophysics
and photochemistry of their prototype [Ru(bipy) ] (bipy =
,2′-bipyridine are well understood. The visible range of the
absorption spectrum is dominated by an intense metal-to-ligand
charge transfer (MLCT) absorption from ruthenium d-orbitals
into the low-lying antibonding π*-orbitals of the bipy ligands
with an absorption maximum at 452 nm and an extinction
coefficient of 14 600 M cm . The UV region is dominated
1
A qualitatively similar picture of the excited-state mecha-
12
nisms is gained for the meridionally coordinated tridentate
2
+
3
2+
2+
analogue of [Ru(bipy) ] , namely, [Ru(tpy) ] (tpy =
2
−5
3
2
2
2
,2′;6′,2″-terpyridine). MLCT absorption occurs at 474 nm
slightly bathochromically and hyperchromically shifted (ε =
−1
−1 13
1
6 100 M cm ) due to the larger accepting π*-orbitals of
3
,6
3
the terpyridine ligands. Upon ISC again MLCT states are
populated.
14,15
In contrast to the bipy counterpart an efficient
−1
−1
4
3
deactivation pathway is available for this emissive MLCT state:
because of the smaller N−Ru−N bite angles in this tpy complex
compared to the parent bipy complex the orbital overlap of the
pyridine nitrogen lone pairs with the ruthenium d orbitals of
the e set (in idealized O symmetry) is lowered. The loss in
4
by π−π* transitions within the aromatic ligands. Following
7
Kasha’s rule rapid vibrational relaxation and internal
conversion occur upon optical excitation populating the
1
lowest-energy MLCT state. From this state nearly quantitative
g
h
8
,9
3
intersystem crossing (ISC) onto the triplet hypersurface
ligand-field splitting shifts d−d excited states ( MC states, MC
3
occurs, which leads to population of the lowest-energy MLCT
1
0
10
state. This state is highly phosphorescent at room
Received: May 27, 2015
11
temperature (λem = 621 nm, ϕ = 0.095 in acetontrile) and
©
XXXX American Chemical Society
A
DOI: 10.1021/acs.inorgchem.5b01151
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
a
Scheme 1. Bis(tridentate)ruthenium(II) and Iridium(III) Polypyridine Complexes
a
2+ 15
2+
+
+
+
20
2+
2+
2+ 21−23
Hammarstro
̈
m (F2+),24 Ruben (G2+),25 Williams (H2+),26 and
By Constable (A ), van Koten (B , I , J , K ), Heinze (C , D , E ),
+
+
from this work (1 and 2 ).
3
22
=
metal-centered) into the energy regime of the MLCT states.
2-ylpyridine-2,6-diamine, ddpd). The increased σ-donor
strength of ddpd compared to tpy sufficiently separates the
These MC states are thermally populated at room temperature
and lead to an efficient emission quenching in [Ru(tpy) ]
λem = 629 nm, ϕ < 1 × 10 ). At 77 K, the available thermal
energy does not suffice to overcome the activation barrier for
2+
3
MC states in the heteroleptic complex [Ru(ddpd)(tpy-
(Scheme 1) from the MLCT states to allow
[Ru(ddpd)(tpy)] to be emissive (ϕ = 0.0045) despite its
2
−5
16
2+
2+
3
(
COOEt)]
D
2+
3
22,23
population of the MC states, and an intense emission is
much lower emission energy (λ = 729 nm).
Gradually
em
13
3
1
regained (ϕ = 0.48).
tuning the vertical MLCT → GS transition energy within a
series of structurally similar [Ru(ddpd)(tpy)]2+ complexes by
variation of appended functional groups decreases the emission
quantum yield with decreasing emission energy following the
energy gap law as pointed out by Meyer and co-workers.
̈
m and co-workers used di(quinolin-8-
yl)pyridine (dqp) as tridentate ligand forming six-membered
Just as this undesirable side effect also the major advantage of
bis(terpyridine)ruthenium(II) complexes over their bipyridine
analogues arises from their coordination geometry. The C
2v
12
30−32
symmetry of the core structure of this complex prevents the
formation of diastereomers even when heteroleptic complexes
bearing ligands with different functional groups are formed.
Syntheses of similarly substituted bipyridine complexes usually
give mixtures of diastereomers that require elaborate methods
Similarly, Hammarstro
2
4,33
chelate rings with ruthenium as metal center.
The
homoleptic complex [Ru(dqp)2]2+
F
2+
(Scheme 1) is
1
7
18,19
to be purified or to be circumvented.
phosphorescent at room temperature (ϕ = 0.02) with a
remarkably long excited-state lifetime of 3.0 μs. Ruben and co-
workers employed the carbonyl analogue of the ddpd ligand,
2,6-di(2-carboxypyridyl)pyridine (dcpp) as chelating ligand
with N−Ru−N bite angles of 90°. The homoleptic complex
Several successful approaches improve the emissive behavior
of bis(tridentate)ruthenium complexes by influencing the
energies of the relevant excited states. Introducing π-
27
accepting functional groups (−SO R, −COOR, Scheme 1,
2
2
+
2+
2+
2+
A , B ) in the ligand backbone (typically in 4′-position)
[Ru(dcpp) ]
G
(Scheme 1) exhibits an extraordinary high
room-temperature emission quantum yield of 0.30 with a long
2
3
lowers the energy of the MLCT states while leaving the energy
3
25
of the MC states unaltered. This hinders the thermal
deactivation process to some extent and increases both
excited-state lifetimes and quantum yields of such compounds
Alternatively, introducing σ-donating
functional groups in the ligand’s periphery directly influences
the energy of the MC states. They are shifted to higher
energies with respect to the MLCT states again hampering
thermal depopulation of the latter. Combining both approaches
yields excited-state lifetimes of up to 50 ns (Scheme 1, A )
excited-state lifetime of 3.3 μs.
Cyclometalation
34,35
(i.e., isoelectronic substitution of a
nitrogen atom for a carbanion in the coordination sphere
around the metal) is discussed as another option for raising the
MC states since the strong σ-donating effect of the anionic
carbon greatly increases the ligand field splitting. While for
−4
15,28
(ϕ ≈ 1−5 × 10 ).
3
3
15
20
3
iridium(III) a large variety of highly phosphorescent cyclo-
3
6−39
metalated complexes are known,
most cyclometalated
2+
15
ruthenium(II) complexes are barely emissive at room temper-
For tris(bidentate)iridium(III) complexes with
cyclometalating ligands of the type [Ir(bipy) (ppy)3−n] (n =
2
+
21,29
40−42
and quantum yields of up to 0.003 (Scheme 1, C )
always at the cost of a lowered excited-state energy.
The MC states are even more efficiently shifted to higher
but
ature.
2
7
n+
n
3
1, 2; ppyH = 2-phenylpyridine) the excited-state mechanisms
energies by widening the N−Ru−N bite angles. This is
that are responsible for the efficient luminescence are well-
43
III
achieved upon introduction of N−CH in between the pyridine
understood. The emissive excited state of Ir complexes is a
3
3
3
rings of the terpyridine ligand (N,N′-dimethyl-N,N′-dipyridin-
linear combination of a mixed MLCT/ LL′CT state (LL′CT =
B
DOI: 10.1021/acs.inorgchem.5b01151
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
+
ligand-to-ligand charge-transfer state) and an energetically very
uncertainty is estimated to be 15%. FD mass spectra were recorded
on a FD Finnigan MAT95 spectrometer. Electrospray ionization
3
similar ligand-centered ( LC) excited state due to the very high
+
+
III
(ESI ) and high-resolution ESI mass spectra were recorded on a
Micromass QTof Ultima API mass spectrometer with analyte solutions
in acetonitrile. Elemental analyses were performed by the micro-
analytical laboratory of the chemical institutes of the University of
Mainz. NMR spectra were obtained with a Bruker Avance II 400
spectrometer at 400.31 ( H) and 100.66 ( C) at 25 °C. Chemical
shifts δ [parts per million] are reported with respect to residual solvent
signals as internal standards ( H, C): CD CN δ( H) = 1.94 ppm,
δ( C) = 1.32 and 118.26 ppm. Electrochemical experiments were
performed with a BioLogic SP-50 voltammetric analyzer at a sample
concentration of 1 × 10 M using platinum wire working and counter
electrodes and a 0.01 M Ag/AgNO reference electrode. Measure-
ments were performed at a scan rate of 100 mV s for cyclic
voltammetry experiments and at 10 mV s for square-wave
voltammetry experiments using 0.1 M [ Bu N][PF ] as supporting
electrolyte in acetonitrile. Potentials are given relative to the
ferrocene/ferrocenium couple (0.40 V vs standard calomel electrode
ligand field splitting of Ir combined with the ppy ligands. This
state is well-separated from other states so that emission
2
6,44−46
becomes very efficient.
Since most cyclometalated
ruthenium complexes are essentially nonemissive at room
temperature in solution it is much more difficult to obtain a
profound understanding of the excited-state mechanisms in
these systems. Berlinguette and co-workers showed that the
1
13
1
13
1
3
13
54
energy gap law is obeyed in complexes of type [Ru(bipy) -
2
+
(
ppy)] demonstrating that direct ISC onto the singlet
−
3
hypersurface followed by vibrational cooling is the dominant
deactivation pathway.
41,42
3
−
1
Van Koten and co-worker recently discussed [Ru(tpy-
−1
+
R)(pbpy-R′)] (pbpyH = 6-phenyl-2,2′-bipyridine) and [Ru-
n
4
6
+
(
tpy-R)(dpb-R′)] (dpbH = 2,6-di(pyrid-2′-yl)benzene) com-
2
0,47,48
plexes and their application in dye-sensitized solar cells.
5
5
Electron-accepting anchor groups (COOR) were appended at
(SCE), E1/2 = 0.90 ± 5 mV under the given conditions). EPR spectra
were recorded on a Miniscope MS 300 X-band CW spectrometer
Magnettech GmbH, Germany). Values of g are referenced against
Mn in ZnS as external standard (g = 2.118, 2.066, 2.027, 1.986,
either one of the two ligands or at both resulting in a series of
+
+
(
weakly or nonemissive tridentate complexes (Scheme 1, I , J ,
2+
+
and K ). A structurally similar weakly emitting iridium(III)
2
+
1.946). Simulations were performed with the EasySpin program
complex (Scheme 1, H ) was synthesized by Williams and co-
5
6
2
6
package. A Horiba LabRAM HR Raman microscope was used for rR
measurements with an object lens (10× NA 0.25) from Olympus.
Samples were optically excited with a red laser (633 nm, 17 mW,
HeNe-laser), green laser (532 nm, 50 mW, air-cooled frequency-
doubled Nd:YAG-solid state laser), or blue laser (473 nm, 20 mW, air-
cooled solid-state laser). Samples were measured in acetonitrile
(Chemsolute, for HPLC) solution in capillary tubes (80 × 1.5 mm,
Marienfeld-Superior).
workers. Despite the fact that the energy gap law is obeyed
within these complex series luminescence quenching is
discussed to arise from thermal depopulation of the very low-
3
3
lying MLCT states via MC states. However, the latter should
be high in energy due to the strong σ-donor strength of the
20
cyclometalating ligand. This apparent discrepancy will be
addressed in this paper.
Density functional theory (DFT) calculations were performed using
In this study we present an extension of our previous work
on tridentate polypyridine ruthenium complexes bearing both
electron-donating amino and electron-withdrawing carboxylic
5
7
the ORCA program package (version 3.0.2). Tight convergence
criteria were chosen for all calculations (keywords TightSCF and
TightOpt, convergence criteria for the SCF part: energy change 1.0 ×
2
+
21,27
acid functionalities (such as A , Scheme 1)
into the field of
−8
−5
−5
1
1
0
E , 1 − E . energy change 1.0 × 10 E , orbital gradient 1.0 ×
h
l
h
cyclometalated complexes and elucidate the excited-state
deactivation mechanisms of these complexes in detail. In the
−5
−7
0 , orbital rotation angle 1.0 × 10 , DIIS error 5.0 × 10 ; for
−6
geometry optimizations: energy change: 1.0 × 10 Eh, maximum
+
+
−4
isomeric [Ru(dpb-R′)(tpy-R)] complexes (1 : R = COOEt, R′
gradient 1.0 × 10 E /bohr, root-mean-square (RMS) gradient 3.0 ×
h
+
−5
−3
=
NHCOMe; 2 : R = NHCOMe, R′ = COOEt) the position of
10
E /bohr, maximum displacement 1.0 × 10 bohr, RMS
h
4
−
displacement 6.0 × 10 bohr). All calculations employ the resolution
of identity (Split-RI-J) approach for the coulomb term in combination
with the chain-of-spheres approximation for the exchange term
the functional groups with respect to the site of cyclometalation
should have a strong impact on their electronic structure and
excited-state ordering. The ground- and excited-state electronic
structures as well as excited-state dynamics are elucidated by a
combination of UV−vis, electron paramagnetic resonance
5
8,59
(
COSX) where Hartree−Fock exchange is required.
optimizations were performed using the GGA functional PBE
combination with Ahlrichs’ split-valence double-ξ basis set def2-SV(P)
for all atoms except ruthenium, which comprises polarization functions
For ruthenium a Stuttgart−Dresden
effective core potential (ECP, def2-SD) was combined with Ahlrich’s
def2-TZVP basis set for the valence electrons.
solvent effects a conductor-like screening model (COSMO) modeling
Geometry
60,61
in
(
EPR), resonance Raman (rR), and emission spectroscopies
6
2,63
and theoretical techniques to provide a better understanding of
the unexplained low-emission efficiencies in cyclometalated
bis(tridentate)ruthenium(II) complexes.
for all non-hydrogen atoms.
6
4,65
To account for
acetonitrile was used in all calculations except for excited-state
gradients. This proved to be particularly important for time-
EXPERIMENTAL SECTION
66
■
General Procedures. Chemicals were obtained from commercial
suppliers and used without further purification. Air- or moisture-
sensitive reactions were performed in dried glassware in an inert gas
atmosphere (argon, quality 4.6). Acetonitrile and dichloromethane
were refluxed over CaH2 and distilled under argon prior to use.
Toluene and xylenes were refluxed over sodium and distilled prior to
dependent DFT calculations (TD-DFT) where gas phase calculations
67
lead to a substantial underestimation of excitation energies. The
optimized geometries were confirmed to be local minima on the
respective potential energy surface by subsequent numerical frequency
analysis (Nimag = 0).
Calculation of EPR parameters and TD-DFT calculations were
performed based on the PBE/def2-SV(P)/ECP(def2-TZVP) opti-
mized geometry of the respective complex employing the triple-ξ basis
set def2-TVZP and several functionals with varying amounts of HF
49
use. Palladium precatalyst [Pd] and the ligand precursors 1-bromo-
2
5
0
51
3
4
,5-dipyridin-2-ylbenzene LA, 4′-chloro-2,2′:6′,2″-terpyridine LB,
52
′-amino-2,2′:6′,2″-terpyridine LC, and ethyl 3,5-dibromobenzoate
5
3
68
69
70
LD were synthesized following literature-known procedures. UV−vis
spectra were recorded on a Varian Cary 5000 spectrometer in 1 cm
cuvettes. Emission spectra were recorded on a Varian Cary Eclipse
spectrometer. Quantum yields were determined by comparing the
areas under the emission spectra on an energy scale recorded for
solutions of the samples and a reference with matching absorbances
exchange: PBE (0%), TPSSh (10%), B3LYP (20%), PBE0
71 72
(25%), and CAM-B3LYP (19−65%). The Douglas−Kroll−Hess
(DKH) relativistic approximation
7
3−76
was used to describe relativistic
effects in these calculations. The DKH keyword in ORCA automati-
7
7
cally invokes adjusted basis sets (TZV_DKH). At least 50 vertical
transitions were calculated in TD-DFT calculations. The electron g
value and hyperfine coupling constants of the unpaired electron to the
11
(
ϕ([Ru(bipy) ]Cl ) = 0.094 in deaerated CH CN). Experimental
3
2
3
C
DOI: 10.1021/acs.inorgchem.5b01151
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
ruthenium atom and all atoms coordinated to ruthenium were
determined in EPR calculations. On the basis of the optimized GS
removed under reduced pressure to give 0.95 g of crude N-acetyl-3,5-
di(pyridin-2-yl)aniline. After column chromatography on neutral Alox
(Brockmann II, 3% water (w/w), solvent ethyl acetate) the product
was obtained as colorless powder. Yield: 848 mg (2.90 mmol, 78%).
Anal. Calcd C H N O (289.33): C, 74.72; H, 5.23; N, 14.52. Found:
1
molecular geometries and the associated Hessian matrices excited-state
gradients were calculated for the lowest 10 excitations at the B3LYP/
def2-TZVP/DKH level of theory to generate excited-state displace-
ments for the rR spectra simulation. The advanced spectra analysis
18
15
3
+
C, 74.93; H, 4.98; N, 14.39%. MS(FD ): m/z (%) = 289.2 (100)
7
8,79
+
+ 1
tool provided with the ORCA program package (orca_asa)
was
[M] , 579.4 (2) [2M+H] . H NMR (CD Cl ): δ [ppm] = 8.67 (ddd,
2 2
3
4
8
4
9
employed to fit the absorption spectra of 1(PF ) and 2(PF ) and to
JHH = 5 Hz, J = 2 Hz, 1 Hz, 2H, H ), 8.38 (t, J = 1 Hz, 1H, H )
6
6
HH
HH
4
2
3
simulate rR spectra. The Gibbs free energy was used to compare
relative energies of the different triplet states of the complexes under
study. Explicit counterions and/or solvent molecules were not taken
into account in all cases. To reduce the computational cost methyl
instead of ethyl groups were used throughout all calculations at the
ester moiety.
8.27 (d, J = 1 Hz, 2H, H ), 8.25 (s, 1H, NH), 7.79 (ddd, J = 8
H
H
H
H
4
5
3
4
6
Hz, J = 1 Hz, 2H, H ), 7.74 (vtd, J = 8, J = 2 Hz, 2H, H ),
HH
HH
HH
3
4
7
7.24 (ddd, J = 8 Hz, 5 Hz, J = 1.2 Hz, 2H, H ), 2.12 (s, 3H
HH
HH
13
1
10
CH ). C{ H} NMR (CD Cl ): δ [ppm] = 169.3 (s, C ), 157.0 (s,
3
2
2
4
8
3
1
6
C ), 150.1 (s, C ), 140.9 (s, C ), 140.0 (s, C ), 137.4 (s, C ), 123.1 (s,
7
9
5
2
11
C ), 121.3 (s, C ), 121.1 (s, C ), 119.2 (s, C ), 24.9 (s, C ).
Synthesis of N-Acetyl-4′-amino-2,2′:6′,2″-terpyridine L2. Proce-
dure (a). XantPhos (123 mg, 213 μmol, 9 mol %) and Pd precatalyst
a
Scheme 2
[
Pd] (39 mg, 53 mmol, 4.5 mol % based on Pd) were dissolved under
2
argon in 15 mL abs. xylenes and left to stand. After 10 min 4′-chloro-
,2′:6′,2″-terpyridine LB (616 mg, 2.30 mmol, 1 equiv), acetamide
151 mg, 2.56 mmol, 1.1 equiv), and sodium tert-butanolate (246 mg,
.56 mmol, 1.1 equiv) dissolved in additional 15 mL of abs. xylenes
2
(
2
were added, and the mixture was heated to reflux for 20 h. Drying of
all reagents prior to use is necessary since the resulting amide is prone
to hydrolysis under the given reaction conditions in the presence of
traces of water. The same workup routine as for N-acetyl-3,5-
di(pyridin-2-yl)aniline was followed. Column chromatography on
neutral Alox (Brockmann II, 3% water (w/w), solvent gradient ethyl
acetate/hexanes 1:3 → 3:1) afforded pure N-acetyl-4′-amino-
2
,2′:6′,2″-terpyridine as off-white powder. Yield: 508 mg (1.75
mmol, 76%). Anal. Calcd C H N O (290.32): C, 70.33; H, 4.86;
N, 19.30. Found: C, 69.92; H, 4.81; N, 19.02%. MS(FD ): m/z (%) =
17
14
4
+
+
+
+ 1
2
90.2 (100) [M ], 313.1 (10) [M + Na] , 603.3 (2) [2M+Na] . H
NMR (0.5 mL of CD Cl + 0.1 mL of deuterated dimethyl sulfoxide):
2
2
2
δ [ppm] = 10.23 (s, 1H, NH), 8.59 (s, 2H, H ), 8.58−8.53 (m, 2H,
H ), 8.48 (d, J = 8 Hz, 2H, H ), 7.76 (vtd, J = 8 Hz, J = 2
Hz, 2H, H ), 7.25 (ddd, J = 8 Hz, 5 Hz, J = 1 Hz, 2H, H ), 2.06
8
3
5
3
4
HH
HH
HH
6
3
4
7
HH
HH
13
1
10
(
1
1
s, 3H, CH ). C{ H} NMR (CD Cl ): δ [ppm] = 170.0 (s, C ),
3 2 2
4
3
8
1
6
56.3 (s, C ), 156.1 (s, C ), 149.0 (s, C ), 148.3 (s, C ), 137.0 (s, C ),
24.1 (s, C ), 121.0 (s, C ), 110.6 (s, C ), 24.5 (s, C ).
7
5
2
11
Procedure (b). To a solution of 4′-amino-2,2′:6′,2″-terpyridine LC
(
2.22 g, 8.94 mmol, 1 equiv) in dichloromethane (30 mL) was added a
solution of acetyl chloride (10 mL, exc.) in dichloromethane (30 mL)
dropwise over a period of 15 min. The resulting mixture was heated to
reflux for 3 h. A slightly yellow precipitate formed during the heating.
After the mixture cooled to room temperature, the solvent and the
acetyl chloride were removed under reduced pressure. The remaining
solid was dissolved in a mixture of water and tetrahydrofuran (1:1, 100
mL), and the pH was adjusted to 8 using aqueous sodium bicarbonate
solution. The resulting colorless precipitate was collected via filtration
yielding microanalytically pure N-acetyl-4′-amino-2,2′:6′,2″-terpyri-
dine L2 (1.09 g, 3.75 mmol). The aqueous phase was further extracted
with dichloromethane (3 × 50 mL). The combined organic phases
were dried over magnesium sulfate and evaporated to dryness. The
crude product was purified via column chromatography on neutral
Alox (Brockmann II, 3% water (w/w), solvent gradient ethyl acetate/
hexanes 1:3 → 3:1) to give a second fraction of pure product as off-
white solid (1.00 g, 3.44 mmol). Yield: 2.09 g (7.20 mmol, 81%). The
a
Buchwald−Hartwig amination of 1-bromo-3,5-di(pyridin-2-yl)-
benzene LA and 4′-chloro-2,2′:6′,2″-terpyridine LB with acetamide
yielding N-acetyl-3,5-di-(pyridin-2-yl)aniline L1 and N-acetyl-4′-
amino-2,2′:6′,2″-terpyridine L2. Atom numbering for NMR assign-
ment is included.
Synthesis of N-Acetyl-3,5-di(pyridin-2-yl)aniline L1. 4,5-Bis-
(
diphenylphosphino)-9,9-dimethylxanthene (XantPhos, 64 mg, 111
μL, 3 mol %) and Pd precatalyst [Pd] (14 mg, 19 μmol, 1.1 mol %
2
based on Pd) were dissolved under argon in 15 mL of abs. toluene and
left to stand. After 10 min 1-bromo-3,5-di(pyridin-2-yl)benzene LA
1
H NMR spectra of both fractions match those obtained from
(
1.15 g, 3.70 mmol, 1 equiv), acetamide (273 mg, 4.62 mmol, 1.25
procedure a).
equiv), and sodium tert-butanolate (444 mg, 4.62 mmol, 1.25 equiv),
dissolved in 15 mL of abs. toluene, were added, and the mixture was
heated to reflux for 8 h. After the mixture cooled to room temperature,
the solvent was removed under reduced pressure, and the remaining
solid was dissolved in concentrated hydrochloric acid (20 mL), water
Synthesis of RuCl (R-tpy), R = COOC H , NHCOCH . A standard
3
2
5
3
procedure was followed for the synthesis of the RuCl (R-tpy)
3
2
9,52
precursors:
Ruthenium(III) chloride hydrate (36% Ru (w/w); R
= COOC H : 1.46 g, 5.20 mmol, 1.3 equiv; R = NHCOCH : 566 mg,
2
5
3
2.01 mmol, 1.3 equiv) was dissolved in ethanol (50 mL), and the
respective terpyridine (R = COOC H : 1.21 g, 3.96 mmol, 1 equiv; R
(
20 mL), and dichloromethane (50 mL). The phases were separated,
and the aqueous phase was extracted twice with dichloromethane (2 ×
0 mL). The aqueous phase was neutralized with dilute aqueous
2
5
= NHCOCH : 450 mg, 1.55 mmol, 1 equiv) was added. The resulting
3
5
mixture was heated to reflux for 3 h during which time the product
precipitated as a red solid. It was filtered off and washed thoroughly
sodium hydroxide solution (pH = 9) followed by extraction with
dichloromethane (3 × 50 mL). The organic fractions of the second
extraction were dried over magnesium sulfate, and the solvent was
with ethanol to remove residual RuCl . The product was dried under
3
reduced pressure and used without further purification. Yield: R =
D
DOI: 10.1021/acs.inorgchem.5b01151
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
a
Scheme 3. Synthetic Procedure
a
Starting from RuCl leading to the heteroleptic cyclometalated ruthenium complex isomers 1(PF ) and 2(PF ) as well as the amino complex
3
6
6
3
(PF ). Numbering for NMR assignments is included.
6
9
13
4B
COOC H : 1.94 g (3.78 mmol, 96%). R = NHCOCH : 730 mg (1.47
(CD CN): δ [ppm] = 215.8 (s, C ), 169.8 (s, C ), 169.3 (s, C ),
3
2
5
3
10
4A
8A
3A
mmol, 95%). Because of the poor solubility of RuCl (tpy-R) no
166.0 (s, C ), 159.8 (s, C ), 155.2 (s, C ), 154.3 (s, C ), 153.1 (s,
3
8B
3B
5B
5A
1B
characterization was performed.
C ), 141.9 (s, C ), 136.9 (s, C ), 136.2 (s, C ), 134.4 (s, C ),
1A
7A
5A
2A
Synthesis of [Ru(dpb-NHCOCH )(tpy-COOC H )](PF ) 1(PF ).
132.5 (s, C ), 127.6 (s, C ), 124.8 (s, C ), 122.9 (s, C ), 122.6 (s,
C ), 120.8 (s, C ), 117.9 (s, C ), 63.3 (s, C ), 24.4 (s, C ), 14.7 (s,
3
2
5
6
6
7B
5B
2B
11
14
RuCl (tpy-COOC H ) (100 mg, 0.195 mmol, 1 equiv) was suspended
3
2
5
12
−1
−1
under argon in 20 mL of abs. acetone, and silver tetrafluoroborate
110 mg, 0.566 mmol, 2.9 equiv) was added. The resulting reaction
C ). IR (KBr disk): λ [cm ] = 3435 (crystal water), 1723 (C
(
Oester), 1711 (CO
), 1600 (CC), 1518 (amide II), 845 (P−F).
max
amide
3
−1
−1
mixture was heated to reflux for 2 h in the dark. After the mixture
cooled to room temperature, the dark brown solution was filtered
through a syringe filter to remove precipitated silver chloride prior to
UV−vis (MeCN): λ (ε) [nm (1 × 10 M cm )] = 241 (49.8),
282 (62.1), 319 (29.8), 378 (14.1), 418 (shoulder, 9.7), 506 (17.3),
555 (13.7).
evaporation of the solvent. The dark, oily residue was dissolved in abs.
Synthesis of [Ru(dpb-COOC H )(tpy-NHCOCH )](PF ) 2(PF ).
2
5
3
6
6
n
BuOH (20 mL), and CH CONH-dpbH L1 (68 mg, 0.234 mmol 1.2
RuCl (tpy-NHCOCH ) (100 mg, 0.201 mmol, 1 equiv) was
3
3
3
equiv) was added. The resulting dark brown to purple solution was
heated to reflux for 16 h giving an intensely colored purple solution.
After removal of the solvent under reduced pressure the remaining
solid was dissolved in acetonitrile (5 mL), and a solution of
ammonium hexafluorophosphate (125 mg, 0.78 mmol, 4 equiv) in
water (1 mL) was added. Addition of more water (∼80 mL) resulted
in the precipitation of the crude product, which was filtered off.
Column chromatography on silica gel (solvent gradient chloroform →
chloroform/methanol 7:1, after a yellow impurity was eluted) afforded
pure [Ru(dpb-NHCOCH )(tpy-COOC H )](PF ) as dark purple
suspended in 20 mL of abs. acetone, and silver tetrafluoroborate
(113 mg, 0.583 mmol, 2.9 equiv) was added. The resulting reaction
mixture was heated to reflux for 2 h in the dark. After the mixture
cooled to room temperature, the dark brown solution was filtered
through a syringe filter to remove precipitated silver chloride prior to
evaporation of the solvent. The dark, oily residue was dissolved in abs.
n
BuOH (20 mL), and C H OOC-dpbH L3 (73 mg, 0.241 mmol, 1.2
2
5
equiv) was added. The resulting dark brown to purple solution was
heated to reflux for 16 h giving an intensely colored red solution. After
removal of the solvent under reduced pressure the remaining solid was
dissolved in acetonitrile (5 mL), and a solution of ammonium
hexafluorophosphate (130 mg, 0.80 mmol, 4 equiv) in water (1 mL)
was added. Addition of more water (∼80 mL) resulted in the
precipitation of the crude product, which subsequently was filtered off.
Column chromatography on neutral Alox (Brockmann II, 3% water
(w/w), solvent gradient chloroform → chloroform/methanol 50:1,
after a yellow impurity was eluted) afforded pure [Ru(dpb-COOEt)-
(tpy-NHCOCH )](PF ) as dark solid. Yield: 87 mg (0.104 mmol,
3
2
5
6
solid. Yield: 114 mg (0.136 mmol, 70%) Anal. Calcd for
C H F N O PRu (839.7)·1.5H O: C, 49.89; H, 3.72; N, 9.70.
3
6
29
6
6
3
2
+
Found: C, 50.01; H, 3.50; N, 9.53%. MS(ESI ): m/z (%) = 347.6 (1)
M-PF ] , 695.1 (100) [M-PF ] , 1535.3 (3) [2M-PF ] . HR-
MS(ESI , m/z): Calcd for C H N O Ru [M-PF ] : 695.1345;
2
+
+
+
[
6
6
6
+
+
3
6
29
6
3
6
1
Found: 695.1336. H NMR (CD CN): δ [ppm] = 9.21 (s, 2H,
3
2
A
3
5A
H ), 8.69 (s, 1H, NH), 8.57 (d, J = 8 Hz, 2H, H ), 8.45 (s, 2H,
HH
5B
2
B
3
3
4
H ), 8.06 (d, J = 8 Hz, 2H, H ), 7.71 (vtd, J = 8 Hz, J = 1
HH
HH
HH
3
6
6A
3
4
6B
Hz, 2H, H ), 7.58 (vtd, J = 8 Hz, J = 1 Hz, 2H, H ), 7.18 (d,
52%) A second fraction consisting of [Ru(dpb-COOEt)(tpy-NH )]-
HH
HH
2
3
8A
7A
3
JHH = 5 Hz, 2H, H ), 7.04−6.95 (m, 2H, H ), 6.93 (d, J = 5 Hz,
(PF ) 3(PF ) was isolated as well (40 mg, 0.050 mmol, 25%). Anal.
HH
6
6
8
B
3
7B
3
2
H, H ), 6.58 (vt, J = 6 Hz, 2H, H ), 4.63 (q, J = 7 Hz, 2H,
Calcd for C H F N O PRu (839.7): C, 51.49; H, 3.48; N, 10.01.
HH
14
HH
12 13
36 29 6 6 3
1
1
1
+
H ), 2.23 (s, 3H, H ), 1.57 (t, J = 7 Hz, 3H, H ). C{ H} NMR
Found: C, 51.46; H, 3.30; N, 9.73%. MS(ESI ): m/z (%) = 347.6 (1)
E
DOI: 10.1021/acs.inorgchem.5b01151
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
1
Figure 1. H NMR spectra of 1(PF ) (upper, blue) and 2(PF ) (lower, red) at room temperature in CD CN (for atom numbering see Scheme 3).
6
6
3
2
+
+
+
[
M-PF ] , 695.1 (100) [M-PF ] , 1535.3 (6) [2M-PF ] . HR-
rigorous protective gas conditions, and only 2.9 equiv of
6
6
6
+
+
MS(ESI , m/z): Calcd for C H N O Ru [M-PF ] : 695.1345;
36
29
6
3
6
Ag[BF ] (instead of 3.6 equiv as usually found in the
4
1
Found: 695.1342. H NMR (CD CN): δ [ppm] = 10.42 (s, 1H,
NH), 9.12 (s, 2H, H ), 8.83 (s, 2H, H ), 8.30 (d, J = 8 Hz, 2H,
20,81
3
literature)
were employed to prevent undesired side
2A
2B
3
HH
reactions. Under these conditions, we obtained the complexes
5
A
3
5B
6A
6B
H ), 8.25 (d, J = 8 Hz, 2H, H ), 7.70−7.60 (m, 4H, H , H ),
HH
3
8B
3
8A
1(PF ) and 2(PF ) in yields of 70% and 52%, respectively,
6
6
7
6
.18 (d, J = 5 Hz, 2H, H ), 7.02 (d, J = 5 Hz, 2H, H ), 6.91−
HH
7A 7B 3
HH
besides small quantities of a side product with a hydrolyzed
acetyl amino group ([Ru(dpb-COOEt)(tpy-NH )](PF )
.82 (m, 2H, H ), 6.76−7.66 (m, 2H, H ), 4.51 (q, J = 7 Hz, 2H,
HH
12 13
1
1
14
3
1
H ), 2.36 (s, 3H, H ), 1.51 (t, J = 7 Hz, 3H, H ). C{ H} NMR
2
6
HH
9B
13
4B
(
1
CD CN): δ [ppm] = 233.7 (s, C ), 171.3 (s, C ), 169.1 (s, C ),
68.7 (s, C ), 159.8 (s, C ), 155.6 (s, C ), 153.7 (s, C ), 152.8 (s,
C ), 145.7 (s, C ), 143.4 (s, C ), 136.5 (s, C ), 136.4 (s, C ),
27.3 (s, C ), 124.5 (s, C ), 124.5 (s, C ), 123.0 (s, C ), 122.7 (s,
3(PF ); see Scheme 3 and Supporting Information).
3
6
10
4A
8A
3A
Characterization of the Isomers. Since both complexes
(PF ) and 2(PF ) are constitutional isomers they share their
8
B
1A
3B
6A
6B
1
6
6
7A
5A
2B
7B
1
elemental composition and show essentially identical mass
spectra and isotope patterns (Supporting Information Figure
S8). The most prominent differences are observed in the NMR
spectra (Figure 1, Supporting Information Figures S9−S18). All
tpy aromatic proton resonances (2A, 5A, 6A, 7A, and 8A)
1
B
5B
2A
11
14
C ), 120.8 (s, C ), 112.8 (s, C ), 61.5 (s, C ), 24.9 (s, C ), 15.0 (s,
1
2
−1
−1
C ). IR (KBr disk): λ [cm ] = 1695 (CO
), 1600 (C
ester,amide
C), 1514 (amide II), 844 (P−F). UV−vis (MeCN): λ (ε) [nm (1 ×
max
3
−1
−1
1
(
0 M cm )] = 242 (49.2), 282 (69.4), 317 (32.4), 351 (16.9), 428
8.7), 502 (15.2), 544 (shoulder, 11.8). (NMR and mass spectrometric
data of 3(PF ) can be found in the Supporting Information.)
1
appear systematically further downfield in the H NMR
6
spectrum (Figure 1, Scheme 3) compared to corresponding
resonances of the dpb protons due to the electron-deficiency of
RESULTS AND DISCUSSION
■
the tpy ligand. Proton resonances of the functional groups
Syntheses of Ru Complexes. Details of the ligand
syntheses can be found in the Experimental Section (L1, L2)
and in the Supporting Information (L3). Several different
experimental protocols have been reported for the synthesis of
heteroleptic ruthenium complexes with terpyridine and
dipyridylbenzene ligands. Complexation can be performed in
+
(
NHCOMe and COOEt) show the same trend. In 1 , with the
tpy-COOEt ligand, the CH and CH resonances of the ethyl
2
3
group are found at 4.63 and 1.57 ppm, respectively, whereas in
+
the regioisomer 2 the corresponding protons of the dpb-
COOEt ligand appear at 4.51 and 1.51 ppm, respectively.
Similarly, the resonance of the acetyl CH protons is found at
water/methanol solution starting from RuCl (tpy-R) at
elevated temperatures and in the presence of a tertiary amine
as sacrificial reductant. This path resembles the microwave-
assisted synthesis that we employed to obtain heteroleptic
bis(terpyridine)ruthenium(II) complexes in high yields.
For cyclometalated complexes, this procedure gives practicable
yields when the coordinating carbon atom is located at one of
the peripheral aromatic rings of a multidentate ligand.
more robust protocol was presented in 1991 by Sauvage and
3
3
.23 ppm in 1+ and at 2.36 ppm in 2 , underlining the
+
2
41
electronic influence of the different ligands. This effect is most
pronounced for the amide NH proton due to its proximity to
the aromatic system in both complexes. The NH resonance of
21,27
+
tpy-NHCOMe in 2 is found at very low field (10.42 ppm),
+
while in 1 , the NH resonance of dpb-NHCOMe is found at
20,41
13
A
8.69 ppm. Similar trends are also observed in the C NMR
spectra (Supporting Information Figures S10 and S15). The
resonance of the carbon atom C9B involved in the metal−
co-workers in the first report on the parent cyclometalated
ruthenium(II) complex [Ru(dpb)(ttpy)] (ttpy = 4′-tolylter-
pyridine): RuCl (ttpy) is activated via chloride abstraction
+
+
carbon bond is found at 215.8 ppm in 1 with the dpb-
8
0
+
NHCOMe ligand and at 233.7 ppm in 2 with the dpb-COOEt
3
with silver tetrafluoroborate. Upon addition of dpbH the
ligand.
desired cyclometalated complex forms readily at elevated
Although the final complexation step requires harsh reaction
conditions and long reaction times, the reaction proceeds in
good yields without significant side reactions. The structural
integrity of the complexes 1 and 2 with all functional groups
is further confirmed by ESI mass spectrometry (Supporting
20,81
temperatures in high yields.
protocol for the synthesis of the complexes presented herein
Scheme 3). Since acetylated amino groups are prone to
hydrolysis and oxidation, all reactions were performed under
We successfully adapted this
+
+
(
F
DOI: 10.1021/acs.inorgchem.5b01151
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
4
−1
Information Figure S8). No significant mass peaks indicating
cleavage of either the ester or the amide are detected.
Interestingly, all mass spectra show a weak peak that can be
10 M cm−1 appears to be a characteristic marker for the
cyclometalation in the central position of the tridentate ligand
and is well-reproduced via theoretical calculations (vide infra).
2
+
assigned to [M−PF ] at m/z = 347.6. This is likely explained
The visible region of the absorption spectra of 1(PF ) and
2(PF ) is dominated by four absorption bands (Tables 1 and
2). The strongest absorption occurs at 506 nm in 1(PF ) and at
502 nm in 2(PF ) and is best described as MLCT transition
involving the pyridine rings of both ligands as acceptors (vide
infra). In complex 1(PF ), the donor and acceptor effect of
the respective ligand and functional group reinforce one
another. An additional absorption maximum appears at the low-
energy side of this MLCT transition at 555 nm. On the
contrary, 2(PF ) only shows a weak shoulder in this region
Figure 2). The overall bathochromic shift of the visible light
absorption features of 1(PF ) compared to 2(PF ) is
accompanied by an increase in absorption intensity. Both
observations are best explained by the increased push−pull
effect arising from the functional groups, which lowers the
highest occupied molecular orbital−lowest unoccupied molec-
ular orbital (HOMO−LUMO) gap (cf. MO diagram,
Supporting Information Figure S22) and increases the
6
6
by the typically rather low oxidation potential of these electron-
rich cyclometalated (polypyridine)ruthenium(II) complexes
6
6
1
(vide infra). IR spectroscopy confirms the presence of both
6
an ester group and a primary amide by characteristic bands for
20
ester CO and amide I stretching vibrations at 1723 and 1711
6
−1
+
−1
+
cm for 1 (DFT: 1733 and 1701 cm , respectively). For 2 ,
the IR stretching vibrations of the ester and amide carbonyl
−1
function are observed at 1697 cm . Indeed, DFT calculations
predict essentially identical vibrational frequencies for both
6
−
1
+
groups at ∼1718−1719 cm for 2 reflecting the different
(
electronic character of tpy and dpb (Supporting Information
6
6
Figure S21). The NH deformation bands are found at ∼1580−
1
600 cm− for both isomers along with coupled C−C
1
vibrations within the aromatic backbone (see rR spectra).
The presence of the PF6 counterion is revealed by broad and
−
−1
intense P−F stretching bands at 840 cm .
UV−vis Spectroscopic Properties of the Cyclometa-
+
+
lated Isomers 1 and 2 . The experimental UV−vis spectra
1
transition dipole moments for the MLCT transitions in
(PF ). In 2(PF ) on the other hand, the donor strength of dpb
and the acceptor strength of tpy are both partially canceled by
of the two complexes 1(PF ) and 2(PF ) are depicted in Figure
6
6
1
6 6
2, and relevant spectroscopic data are summarized in the
the substituents leading to a larger HOMO−LUMO gap and
1
weaker MLCT absorptions. As a consequence, 1(PF ) appears
6
black in the solid state and deep purple in acetonitrile solution,
while 2(PF ) is dark red in the solid state and in solution. This
6
observation is in accordance with previous results for similar
20
complexes.
Both compounds show two more absorption bands between
1
3
40 and 450 nm with similar intensities as the MLCT
absorptions. In the literature the origins of these bands are
consistently discussed as MLCT transitions involving the
20,41,81
cyclometalating ligand.
However, this interpretation
contradicts the observed hypsochromic shift of this absorption
from 378 nm (1(PF )) to 351 nm (2(PF )) upon exchanging
6
6
the electron-donating N-acetyl amino group for an electron-
accepting carboxy group. A more consistent explanation of
these absorptions is gained from TD-DFT calculations and rR
experiments.
Theoretical calculations were performed using geometries
optimized at the PBE/def2-SV(P) level of theory with an
effective core potential at ruthenium (def2-SD, def2-TZVP).
This level of theory for the geometry optimization was chosen
based on data obtained by screening multiple functionals and
basis sets and by comparison to geometrical parameters
obtained from crystal structures of three structurally related
compounds ([Ru(dpb-COOMe)(tpy)](PF ), [Ru(dpb)(tpy-
6
20
Figure 2. Experimental absorption spectra of 1(PF ) (upper, blue) and
COOEt)](PF ), and [Ru(pbpy-COOMe)(tpy)](PF )). This
6
6
6
−5
2
(PF ) (lower, red) in acetonitrile at room temperature; c = 2 × 10
mol l and calculated UV−vis spectra from TD-DFT calculations
B3LYP, black). The calculated spectra are shifted by 1000 cm to
evaluation showed only a marginal dependence of the
geometrical parameters on the size of basis set allowing for
usage of a rather small and computationally efficient basis set.
The variation of the structural parameters with the choice of
functional was larger, but still, qualitatively similar results were
obtained with all functionals under study (BP, PBE, BLYP,
TPSS, TPSSh, M06L, B3LYP, PBE0). Remarkably, within the
mean deviation the hybrid functionals PBE0 and B3LYP
yielded identical optimized geometries compared to the
corresponding GGA functionals PBE and BLYP. Hence, the
more economic GGA functional PBE was preferred over hybrid
functionals for geometry optimizations.
6
−
1
−1
(
lower energies to match calculated and experimental π−π* absorption
energies.
Experimental Section. Both complexes show absorption
features of very similar shape, intensity, and energy between
2
00 and 325 nm. This is because of the identical ligand
backbone of both complexes and is characteristic for [Ru-
+
20,81
(dpb)(tpy)] -type complexes.
Especially the sharp absorp-
tion band at ∼320 nm with an extinction coefficient of ∼3.0 ×
G
DOI: 10.1021/acs.inorgchem.5b01151
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Table 1. Spectral Decomposition of the Absorption Bands of 1(PF ) in Acetonitrile Solution in the Range from 9000 to 29 000
6
−1
a
cm Using the Advanced Spectral Analysis Tool of ORCA (orca_asa)
experimental data
theoretical data
state
λ−1, cm−1
λ, nm
fosc
εmax, M−1 cm−1
state
2
λ−1, cm−1
fosc
assignment
1.39 × 10−
1.15 × 10−
1.48 × 10−
0.102
3
2
2
145
Ru→tpy MLCT
3
1
2
3
4
12 935
15 438
16 197
17 711
773
648
617
565
5.5 × 10−3
Ru→tpy MLCT
1
714
15 095
1940
11 300
4
5
18 194
19 150
0.030
0.040
Ru→tpy
1MLCT
6.99 × 10−
0.107
3
1570
5
6
7
18 694
19 687
20 910
535
508
478
1
11 700
6200
6
7
20 914
21 962
0.091
0.213
Ru→dpb MLCT
0.124
Ru→tpy
1MLCT
4.13 × 10−
5.38 × 10
7.88 × 10−
3
2
3
690
8
21 843
2.6 × 10−3
Ru→dpb MLCT
1
8
9
21 399
23 871
25 166
26 500
467
419
397
377
−
5290
1210
14 700
1
1
0
1
0.234
a
−1
Eleven Gaussian bands were required to reproduce the shape of the absorption profile, and λ , f , and ε are obtained from the respective fitted
osc
max
bands. Theoretical data are obtained from TD-DFT calculations (B3LYP) and assigned to the experimental bands based on their energy and
oscillator strengths.
Table 2. Spectral Decomposition of the Absorption Bands of 2(PF ) in Acetonitrile Solution in the Range from 9000 to 29 000
6
−1
a
cm Using the Advanced Spectral Analysis Tool of ORCA (orca_asa)
experimental data
theoretical data
state
λ−1, cm−1
λ, nm
fosc
εmax, M−1 cm−1
state
λ−1, cm−1
fosc
assignment
3
3
2
3
1
2
3
4
15 479
17 345
17 455
19 445
646
577
573
514
1.29 × 10−
1.49 × 10−
3.25 × 10−
0.216
208
Ru→tpy MLCT
1
331
4
5
7
8
18 332
18 593
21 675
21 681
0.014
0.060
0.170
0.096
Ru→tpy MLCT
1
2530
13 000
Ru→tpy MLCT
1
Ru→dpb MLCT
2
3
2
5
6
7
20 269
21 472
23 244
493
466
430
1.54 × 10−
9.72 × 10
7.42 × 10−
2460
1430
6240
−
11
16
24 120
26 366
0.111
0.026
Ru→tpy,dpb
1MLCT
4
2
2
8
9
24 575
25 035
28 059
29 114
407
399
356
343
1.39 × 10−
1.12 × 10−
1.16 × 10−
0.504
56
1
1480
1250
16 300
Ru→tpy MLCT
1
0
1
1
a
−1
Eleven Gaussian bands were required to reproduce the shape of the absorption profile, and λ , f , and ε are obtained from the respective fitted
osc
max
bands. Theoretical data are obtained from TD-DFT calculations (B3LYP) and assigned to the experimental bands based on their energy and
oscillator strengths.
Vertical excitations were generated within the TD-DFT
formalism with a triple-ξ basis set (def2-TZVP) and the
Douglas−Kroll−Hess relativistic approximation in combination
with functionals of varying Hartree−Fock (HF) exchange
match reasonably well in the UV region, when the entire
theoretical spectrum is shifted by 1000 cm− to lower energies
(Figure 2). In the visible region, however, the agreement is
somewhat lower for both 1(PF ) and 2(PF ). This is mostly
1
6
6
(
PBE, 0%; TPPSh, 10%; B3LYP, 20%; PBE0, 25%). Addition-
because of the weakness of TD-DFT in the description of
6
8,82,83
ally, the range-separated CAM-B3LYP functional was em-
ployed. While PBE and CAM-B3LYP both gave unsatisfactory
results, TPSSh, B3LYP, and PBE0 performed equally well in
TD-DFT calculations of 1(PF ) and 2(PF ) compared to the
charge transfer excitations.
10 lowest excitations and all further significant transitions (f >
Difference density plots of the
+
+
0.01) are shown for 1 and 2 in the Supporting Information,
Tables S1 and S2. In both complexes the lowest-energy
transition is a well-defined HOMO−LUMO excitation with no
significant admixture of other orbitals. The LUMO in both
complexes is a π*-orbital located at the tpy ligand, while the
HOMO is essentially a π-orbital of the dpb ligand. The lowest-
energy excitation can formally be regarded as a LL′CT
transition although ruthenium d-orbitals of the t set are
6
6
corresponding experimental data (see Supporting Information,
Figures S23 and S24). A systematic increase of all transition
energies and transition probabilities (oscillator strengths) was
observed with increasing HF exchange (TPSSh < B3LYP <
PBE0).
In the following the TD-DFT calculations using B3LYP as
2
g
functional and the corresponding spectra generated with
orca_mapspc (line width 1500 cm ) will be discussed and
correlated with the experimental absorption spectra of 1(PF6)
admixed into the frontier orbitals (mixed LL′CT and MLCT
character). To differentiate between these transitions and the
MLCT transitions these states will be labeled LL′CT. Since
both ligands are perpendicular to each other so are the
−1
and 2(PF ). The theoretical and experimental UV−vis spectra
6
H
DOI: 10.1021/acs.inorgchem.5b01151
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
contributing d-orbitals, which renders the metal contribution
negligible. As a consequence, these HOMO−LUMO tran-
+
+
sitions for 1 and 2 are symmetry-forbidden excitations due to
−
5
the vanishing overlap integral (oscillator strengths of 1.1 × 10
and 7.1 × 10− for 1+ and 2 , respectively) and do not
contribute to the absorption spectrum. The corresponding
LL′CT states, however, might play a significant role for the
excited-state behavior of 1(PF ) and 2(PF ), as discussed
5
+
3
6
6
below.
The higher-energy excitations predicted in the visible region
1
are MLCT transitions from metal d-orbitals onto the ligands.
Interestingly, these excitations do not only involve tpy π*-
orbitals. Already in the range above 400 nm dpb accepting
orbitals play a major role for the absorption profile. The UV
transitions in the range between 400 and 320 nm also consist of
1
MLCT transitions onto both ligands. A distinct separation of
Ru → tpy and Ru → dpb MLCT transitions, with the former
being responsible for the low-energy absorption band between
6
3
00 and 450 nm and the latter yielding the UV band between
20 and 430 nm, is not valid.
To be able to match the experimentally obtained spectrum
with the theoretical data spectral decompositions of the visible
range of the absorption spectra were performed. This is of
particular interest considering the predicted low intensity of the
LL′CT transitions in 1(PF ) and 2(PF ). The fit data are
6
6
summarized in Tables 1 and 2 for 1(PF ) and 2(PF ),
6
6
respectively. Despite the high quality of the fit (see Supporting
Information, Figure S25) the low-intensity LL′CT absorptions
in the low-energy edge of the absorption spectrum were not
detected in either case. An upper limit of the oscillator
−4
strengths of these transitions is estimated as f ≤ 1 × 10 , which
is in agreement with the computational data. The decom-
position of the absorption bands supports the qualitative
discussion of the spectra above. A plausible assignment of TD-
DFT excitations to the most intense bands was possible based
on similarities in oscillator strengths and transition energies
underlining that the computational method gives a reasonable
estimate of the absorption spectrum. The lowest-energy
excitation observed within the spectral decomposition
−1
(
12 935 and 15 479 cm for 1(PF ) and 2(PF ), respectively)
6 6
could not be assigned to any calculated vertical singlet
3
excitation. We ascribe these to MLCT absorptions that
become partially spin-allowed due to spin−orbit coupling in
4
,5
the presence of ruthenium. A complete assignment of all
observed bands is of course out of reach at the presented level
of theory and will generally be very difficult based on the
complexity of the absorption characteristics of 1(PF ) and
6
Figure 3. (a) Resonance Raman spectra of 1(PF ) in acetonitrile
6
2
(PF6).
Resonance Raman Studies on 1 and 2 . To further
support this interpretation of the absorption characteristics of
(PF ) and 2(PF ) in the visible region rR spectroscopic
−1
solution (298 K) in the range of 1400−1700 cm at different
+
+
excitation wavelengths. (b) Experimental (blue) and DFT-calculated
−1
(
black, line width = 10 cm ) rR spectra of 1(PF ) at 473, 532, and
6
1
633 nm excitation wavelength. Asterisks indicate Raman bands of
6
6
studies in acetonitrile solution were performed. This technique
has proven to be useful just recently in the elucidation of the
charge redistribution upon optical excitation in bis-
CH CN.
3
spectra obviously directly depends on the extinction coefficient
at the given irradiation wavelength so that a maximum in rR
intensity is expected in the range of 530−470 nm for both
complexes. On the other hand, the rR intensity depends on
whether a given vibrational mode contributes to the geo-
metrical reorganization associated with the given optical
transition at the Franck−Condon point. Remarkably, the
carbonyl stretching vibrations of both the ester and the
amide functionality only play a subordinate role at all excitation
wavelengths for both 1(PF ) and 2(PF ). Since the common
84,85
(terpyridine)ruthenium(II) complexes.
The rR spectra of
1
(PF ) and 2(PF ) with excitation at 473, 532, and 633 nm are
6
6
shown in Figures 3 and 4, respectively. Since even the idealized
core symmetry of these complexes (C ) is rather low and the
number of atoms is high (N = 76), a multitude of Raman
bands, many of them overlapping, is observed in the rR spectra.
Qualitatively, the spectra appear very similar at the different
excitation wavelengths with the only differences lying in the
intensities of the bands. On one hand, the intensity of the rR
2
v
6
6
I
DOI: 10.1021/acs.inorgchem.5b01151
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
stretching bands are observed at any excitation wavelength
(
cf. IR spectrum, vide supra).
Since the MLCT transitions involve π*-orbitals of the
aromatic ligands the corresponding CC valence vibrations
should be excited and give intense rR responses (see Figures 3a
and 4a). They are assigned based on DFT-calculated ground-
state vibrational frequencies. The bands at 1600 cm− for
1
1
(PF ) and 2(PF ) are assigned to the symmetric valence
6 6
vibrations (local A symmetry) of the aromatic rings. While for
1
1
(PF ), all these vibrations are very close in energy (DFT:
6
−1
between 1597 and 1605 cm ) and overlap in the rR spectra at
all excitation wavelengths, the symmetric vibration of the
carboxy-substituted phenyl ring of 2(PF ) is shifted by 20 cm
−1
6
to lower energy yielding a well-resolved band in the rR spectra
−1
−1
of 2(PF ) at 1581 cm (calculated at 1584 cm ) that is not
6
present in 1(PF ). The significant participation of the phenyl
6
vibration in the rR spectra of 2(PF ) indicates substantial Ru→
6
dpb character of the MLCT absorption band at 532 nm and at
4
73 nm. This corroborates the findings of TD-DFT calculations
that Ru→dpb excitations are present even at energies below
50 nm. Since the corresponding vibration of the phenyl
4
moiety of 1(PF ) overlaps with those of the pyridine rings a
6
similar conclusion cannot be drawn for 1(PF ) although the
6
width of the band suggests participation of all six totally
symmetric aromatic vibrations. Similar behavior is observed for
the antisymmetric (local B symmetry) CC valence vibrations
2
−1
at ∼1525 cm . While for 1(PF ) these vibrations overlap
yielding one broad Raman band, a distinct shoulder at 1514
cm appears for 2(PF ), which is assigned to the phenyl
6
−1
6
moiety again underlining the mixed Ru→tpy/Ru→dpb MLCT
character of the absorption band at 473 and 533 nm.
The intensity of the bands between 1400 and 1550 cm−
increases substantially upon increasing λexc from 473 to 532 nm.
Ground-state vibrational frequencies and Raman intensities
provide no straightforward explanation for that. The
independent mode-displaced harmonic oscillator (IMDHO)
model was employed, which assumes harmonic ground- and
excited-state potential energy surfaces (PES) and no frequency
changes upon excitation. Additionally, the excited-state PES are
considered as displaced with respect to the ground-state PES
along certain (or all) normal modes. More evolved theoretical
approaches have been employed previously in the description
1
86
of rR spectra of large molecules, but these require much more
computational time and are limited to a small number (2−3) of
electronically excited states that can be considered in the
calculations.
At least three prerequisites must be met to yield a
qualitatively good description of the rR behavior of a given
compound: First, a high quality of the normal mode
displacements is crucial for a reasonable description of the rR
spectra since they determine the intensity of the corresponding
Raman bands. These can be computed within the above-
mentioned theoretical model from excited-state gradient
Figure 4. (a) Resonance Raman spectra of 2(PF ) in acetonitrile
6
−1
solution (298 K) in the range of 1400−1700 cm at different
excitation wavelengths. (b) Experimental (red) and DFT-calculated
−1
(
black, line width = 10 cm ) rR spectra of 2(PF ) at 473, 532, and
6
6
33 nm excitation wavelength. Asterisks indicate Raman bands of
CH CN.
3
78,87
calculations.
Second, the vibrational frequencies obtained
description of the intense visible range absorption band is that
of a MLCT transition onto the terpyridine ligand the
from DFT calculations must correspond well to the
experimentally observed ones since these define the normal
modes and have a large impact on the displacement parameters.
Third, the character of the calculated electronic excited states
must match that of the actual transitions. This is the most
challenging part, especially for charge transfer processes, since
−1
terpyridine ester CO vibration at 1723 cm in 1(PF6)
should be visible for all excitation wavelengths under study. At
λexc = 633 nm this is indeed true, but at higher energies (λexc
32 nm) the contribution of this vibration diminishes until
=
5
68,82,83
disappearance at λexc = 473 nm. At shorter wavelengths, Ru→
DFT has its weakness in describing such excitations.
dpb MLCT excitations become increasingly relevant for the
All calculations were performed based on the B3LYP/def2+-
absorption characteristics. For 2(PF ), no distinct CO
SV(P)/DKH/COSMO(acetonitrile) optimized geometry of 1
6
J
DOI: 10.1021/acs.inorgchem.5b01151
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
+
and 2 . Vibrational frequencies were obtained at the same level
prevail (excitations 4 and 5), the absorption at 533 nm is
dominated by excitation 5 (Ru→tpy MLCT; see Table 2 and
Supporting Information Table S2). At λexc = 473 nm Ru→dpb
transitions come into play (state 7). Again, charge transfer
processes into the electron-rich dpb ligand occur at significantly
lower energies than expected. In essence absorption bands
between 550 and 450 nm consist of MLCT absorptions from
of theory. Since ORCA does not support a solvation model for
excited-state gradients, the 10 lowest vertical excitations and
gradients were generated in the gas phase at the B3LYP/def2-
TZVP/DKH level of theory. These 10 excitations describe the
spectral range under study sufficiently well. The orca_asa
software was then used to simulate the first-order rR spectra
(higher-order Raman bands were considered but did not
ruthenium onto both ligands in both complexes 1(PF ) and
6
improve the quality of the simulations). A homogeneous line
2(PF6).
−1
Emission Properties of 1+ and 2 . Cyclometalated
polypyridine complexes of ruthenium usually exhibit only
weak emission. The carboxy-substituted complex [Ru(dpb)-
+
broadening of 1200 cm was assumed for all 10 excitations.
The vibrational frequencies of both compounds were uniformly
scaled by a factor of 0.967. This factor yields a maximum of
agreement between the experimental and simulated spectra.
Neither the different MLCT optical excitations nor the
molecular vibrations are energetically separated. Consequently,
the rR spectrum at a given excitation wavelength is a
superposition of rR profiles of the individual electronic
excitations weighted by their contribution to the absorption
spectrum at that wavelength. At the same time the individual
Raman bands are a superposition of multiple vibrational modes.
Despite the large number of approximations and assumptions
+
(tpy-COOR)] is nonemissive at room temperature, while its
+
regioisomer [Ru(dpb-COOR)(tpy)] shows weak room+-
20
temperature emission. Similarly, [Ru(pbpy)(tpy-COOR)]
20
is nonemissive at room temperature.
2(PF ) emits at room temperature at 751 nm with an
6
−5
emission quantum yield of 1.4 × 10 (Supporting Information
Figure S26). At 77 K, a much more intense emission is
observed with an emission maximum at 716 nm and a band
shape typical for a ruthenium-based emission arising from a
single vibronic progression (see Supporting Information Figure
the experimental rR spectra of 1(PF ) are remarkably well-
6
27,88
reproduced by these simulations. This allows further con-
firmation of the character of respective absorption bands by
assigning optical transitions to the respective absorption
S27).
The temperature dependence of the phosphorescence of
2(PF ) is shown in Figure 5. The emission intensity rapidly
6
energies. The shape of the rR spectrum of 1(PF ) at 633 nm
6
is dominated by the most intense low-energy optical transitions
4
and 5 (see Table 1 and Supporting Information Table S1),
which are Ru→tpy MLCT transitions. At λexc = 532 nm, the
character of the involved absorptions changes and so does the
−
1
rR spectrum. The range of 1000−1300 cm (in-plane
deformation vibrations of the ligand backbones) is very
characteristic for these changes, and a good agreement between
simulation and experiment is obtained. The rR spectrum at λexc
=
532 nm is dominated by the optical excitations 6 and 7 (Ru→
tpy and Ru→dpb; see Table 1 and Supporting Information
Table S1). Consequently, the absorption region around the
MLCT absorption maximum is composed of both Ru→tpy and
Ru→dpb MLCT transitions. This is in contrast to the widely
accepted picture of this band as exclusively arising from Ru→
tpy transitions. The quality of the rR simulation decreases
somewhat at λexc = 473 nm. This is most likely because, at this
wavelength, absorption bands of the second MLCT absorption
Figure 5. Emission quantum yield of 2(PF ) in fluid butyronitrile
6
solution in the temperature range between 160 and 300 K (lower to
−1
upper). (inset) Plot of ln(ϕ) vs T and the corresponding linear fit
0
curve based on ln(ϕ) = ln(k /k′ ) + ΔE′/R·1/T (see text for
r
(between 450 and 320 nm) contribute already. To keep the
explanation).
computational effort manageable these electronic transitions
were neglected and thus are missing in the simulation.
Consequently, the number of predicted rR bands is lower
than experimentally observed. It is worth noting that the
calculations also give an explanation for the missing resonant
increase of the carbonyl stretching vibrations: For almost all
major optical excitations the CO fragment is located in a
nodal plane of the involved orbitals and thus is only marginally
affected by the rR effect.
increases upon lowering the temperature. This behavior is easily
understood following the argumentation of van Houten and
2
3
Watts and later Meyer and co-workers. The lifetime of the
3
emissive MLCT state depends on the rates of radiative and
nonradiative decay, k and knr. Additionally, irreversible thermal
r
3
3
depopulation of the emissive MLCT states via MC states is a
relevant nonradiative relaxation pathway in (polypyridine)
ruthenium complexes. Because of the irreversibility of this
2
(PF ) gives a qualitatively very similar picture although the
6
0
process it can be accounted for by a third rate constant k′ and
overall agreement between experiment and simulation is
an Arrhenius-like activation barrier ΔE′. As the quantum yield
slightly lower (Figure 4). Especially, the B symmetric CC
2
3
−1
is proportional to the lifetime of the emissive MLCT state and
valence vibrations at 1525 cm seem to be missing in the
3
−1
the rate constant for radiative decay, the following relationship
simulation as they are calculated at higher energy at 1556 cm .
At 632 nm excitation wavelength essentially off-resonance
excitation is achieved leading to a spectrum with a low signal-
to-noise ratio. Still, the absorption characteristics at the
different wavelengths are identical. While at 632 nm low-
energy MLCT transitions from ruthenium to the tpy ligand
between ϕ and T is obtained:
0
ϕ = k /[k + k + k′ ·exp(−ΔE′/RT)]
r
r
nr
This equation gives a nonlinear relationship between ln(ϕ)
−
1
3
and T , as has been shown by Meyer and co-workers. In the
K
DOI: 10.1021/acs.inorgchem.5b01151
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
+
+
Figure 6. Jablonski diagrams and electronic spin densities of B3LYP-optimized triplet states of 1 and 2 (contour plots at 0.001 isosurface value).
1
1
3
+
MLCT and LL′CT energies are obtained from TD-DFT calculations. LL′CT energy of 1 is obtained as energy difference from relaxed singlet and
3
+
triplet geometries from DFT calculations. The MLCT energy of 2 is determined experimentally from the E emission at 77 K. Other triplet-state
00
energies are obtained from B3LYP geometry optimizations.
present case, however (see inset of Figure 5), a linear
C−C−N
dihedral angles of ∼9°). This nicely illustrates
peripheral
−
1
relationship between ln(ϕ) and T is obtained in the
temperature range from 180 to 300 K. This can be explained
assuming an efficient irreversible excited-state deactivation via
the dissociative character of this excited state that has
previously been illustrated for a series of other bis(tridentate)
23
ruthenium complexes and that is responsible for the intrinsic
photochemical reactivity of (polypyridine)ruthenium com-
3
low-lying MC states with a small barrier ΔE′. With this
2
,3,15
3
3
assumption and at sufficiently high temperatures k and k
plexes.
The MC− MLCT energy difference is determined
r
nr
−1
−1
become negligible with respect to the exponential term
as 9.4 kJ mol and 19.0 kJ mol (B3LYP and PBE0,
respectively). Even though this energy difference is not directly
related to the experimentally determined activation barrier ΔE′
0
associated with the rate constant k′ , and ln(ϕ) indeed depends
−1
linearly on T . From the linear regression, the thermal
3
3
−1
activation barrier for the MLCT− MC surface crossing was
of 11.4 kJ mol it serves as a lower limit to the latter. The
−1
determined to be ΔE′ = 11.4 ± 0.5 kJ mol , which is just a
calculation using B3LYP as functional seems to give a good
2
+
3
3
fourth of the activation barrier found for [Ru(bipy) ] (ΔE′ =
estimate to the MLCT− MC energy difference, while PBE0
3
−
1
3
4
2.6 kJ mol ).
To get a clearer picture of the involved excited-states, DFT
calculations were performed. Using B3LYP or PBE0 as
slightly overestimates this energy gap.
3
3
+
The MLCT− MC energy difference calculated for 2 is
significantly lower than that obtained for [Ru(tpy-COOH)(tpy-
3
3
2+
−1 23
functional, one MLCT and one MC state could be localized
for 2 (see Figure 6). The geometry of the MLCT state is
essentially unaltered compared to the GS geometry. Because
NH )] at a similar level of theory (26.8 kJ mol ). This
2
+
3
might be attributed to the fact that the strong σ-donating effect
of dpb is partially diminished by the electron-accepting ester
functionality in 4-position and the tpy ligand is a weaker
electron acceptor. More importantly though, cyclometalation of
1
of the dipolar character of this excited state the electron-
deficient ruthenium atom is slightly shifted by 3 pm toward the
tpy ligand, while the dpb−tpy distance remains unaffected
corresponding to a simple motion of ruthenium toward tpy in
3
the central phenyl ring only raises one of the three MC states
that are responsible for the excited-state deactivation, while the
3
the fixed N C coordination sphere. A much stronger distortion
other MC states remain low in energy and are efficiently
5
1
3
20
with respect to the GS geometry is observed for the MC state
populated at room temperature.
since an antibonding metal orbital is occupied. The Ru−N
In contrast to 2(PF ), 1(PF ) is nonemissive both at room
tpy
6
6
bond lengths are substantially elongated by ∼20 pm, and the
peripheral pyridine rings of the tpy ligand are significantly
temperature and at 77 K. This cannot be accounted for with a
thermally activated deactivation mechanism of the excited state
unless the activation barrier is close to zero. Hence another
twisted out of the plane of the central pyridine ring (Ncentral
−
L
DOI: 10.1021/acs.inorgchem.5b01151
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
mechanism must be responsible for the excited-state
deactivation. As was shown above by TD-DFT calculations,
symmetry-forbidden LL′CT transitions exist at the low-energy
edge of the absorption spectra of 1(PF ) and 2(PF ). DFT
6
6
3
calculations yielded a corresponding LL′CT excited state for
+
3
3
1
(see Figure 6) as well as MLCT and MC states with spin
+
distributions similar to those of 2 . Remarkably, the complex
3
+
core is essentially undistorted for the LL′CT state of 1
although the Ru−C bond is slightly shortened by 3 pm
dpb
and the central Ru−N is elongated by 6 pm corresponding to
tpy
the movement of ruthenium toward the dpb ligand within a
3
+
fixed ligand framework. The MLCT state of 1 , however, is
1
+
slightly distorted compared to the GS geometry of 1 : The
central pyridine ring of the tpy ligand is somewhat shifted out
of the plane perpendicular to the dpb ligand. Again, the
ruthenium atom is closer to the tpy ligand because of the
dipolar character of this excited state. The dissociative character
3
+
of the MC state is also found for 1 with characteristically
elongated Ru−N bond lengths and a significant distortion of
tpy
the peripheral pyridine rings away from the metal center. For
Figure 7. Cyclic voltammograms of 1(PF ) (upper, blue) and 2(PF )
6
6
+
3
emissive 2 , the LL′CT state could not be found. However, the
(lower, red) (c = 1 mM) in 0.1 M acetonitrile solution of
n
1
[
Bu N][PF ] at 298 K. Potentials are referenced against the FcH/
corresponding LL′CT absorption is calculated to be ∼3800
4 6
+
−
1
−1
+
FcH couple (E = 0.40 V vs SCE).
cm (45 kJ mol ) higher in energy as compared to that of 1
1/2
3
(
Figure 6). Hence, we suggest that the LL′CT state does not
play a significant role for the excited-state dynamics of 2 .
+
+
+
Table 3. Electrochemical Data of 1 , 2 , and
+
n
[
(dpb)Ru(tpy)] (1 mM) in 0.1 M [ Bu N][PF ] Electrolyte
4 6
a
The order of these three states gives a straightforward
Solution
explanation to the nonemissive behavior of 1(PF ). The lowest-
6
3
lying triplet excited state is the LL′CT state (Figure 6). Similar
Eox,1, V Eox,2, V Ered,1, V Ered,2, V Ered,3, V
1
to the corresponding LL′CT transition in the absorption (vide
1+
2+
0.02
0.14
0.12
0.77
1.28
1.36
−1.82
−1.96
−1.95
−2.25
−2.29
−2.42
3
1
supra), a LL′CT→ GS emission process is symmetry-
forbidden due to the orthogonality of the two ligands. The
only available deactivation pathway is via radiationless ISC into
the ground state followed by vibrational relaxation. The
b
[
Ru(dpb)(tpy)]+
a
+
Potentials are given in volts and referenced against the FcH/FcH
b
20
couple (E1/2 = 0.40 V vs SCE). Values taken from the literature.
3
MLCT state that could evolve into the ground state radiatively
is ∼10 kJ mol higher in energy than the LL′CT state and is
only very inefficiently populated at room temperature.
−
1
3
II
III
to the Ru /Ru redox couple with contributions from the
3
Apart from the LL′CT state described by the strongly
highest occupied π-orbital of the cyclometalating ligand
+
+
electron-accepting tpy and the electron-donating dpb ligand,
(HOMO of 1 and 2 , Supporting Information Figure S22)
3
3
the MLCT− MC separation is calculated to be substantially
as evidenced from Mulliken spin population analysis and spin
+
+
2+
2+
larger in 1 than in 2 also due to the stronger push−pull
density plots of 1 and 2 (Supporting Information Figures
S30 and S31). The origin of the second oxidation process,
however, is less clear. DFT calculations suggest a mixed
oxidation of the metal center and cyclometalating ligand as
primary step (see Supporting Information Figure S32), but the
irreversibility of this process points to follow-up reactions, so
+
substitution in 1 that stabilizes the tpy-based LUMO while at
the same time destabilizing the metal-centered excited states.
In summary, the introduction of a carbon atom in the
coordination sphere of ruthenium indeed increases the
3
3
MLCT− MC energy gap sufficiently to render [Ru(dpb)-
+
+
II
(
(
tpy)] and 2 emissive at room temperature, while [Ru-
that the final oxidation product remains unidentified. The Ru /
tpy)2]2 is silent. This effect is reinforced by attaching
+
III
Ru oxidation is shifted by 0.6−0.7 V to lower potentials
additional donor and acceptor functionalities in the ligand
compared to bis(terpyridine)ruthenium complexes bearing the
3
3
89
periphery that further increase this MLCT− MC energy gap.
same functional groups. This illustrates the strong σ-donor
Unfortunately, this push−pull approach suffers from the
character of the cyclometalating ligand that greatly increases the
electron density at the metal center. The NHCOMe group at
3
concomitant formation of a very low-lying LL′CT dark state
when the donor−acceptor substitution becomes too strong.
Electrochemical Properties of 1(PF ) and 2(PF ). The
the cyclometalating ligand in 1(PF ) indeed leads to a further
6
shift of the ruthenium-based oxidation by 0.10 V to lower
6
6
+
spatial orientation and symmetry of the frontier orbitals were
further experimentally probed by electrochemical and EPR
values as compared to [Ru(tpy)(dpb)] , while the COOEt
substitution of the dpb ligand in 2(PF ) slightly increases this
6
experiments. The cyclic voltammograms of the complexes
redox potential by 0.02 V.
n
1
(PF ) and 2(PF ) in acetonitrile using 0.1 M [ Bu N][PF ] as
In the range accessible for reduction (up to −2.5 V vs FcH/
6
6
4
6
+
supporting electrolyte are shown in Figure 7 and Supporting
Information Figures S28 and S29. Both complexes show a+
FcH ) three quasireversible or irreversible redox waves are
found for 1(PF ), while for 2(PF ) only two reduction waves
6
6
reversible oxidation wave at a potential of 0.02 V vs FcH/FcH
are detected. These are assigned to ligand-centered reductions
with the first one centered on the tpy ligand. The localization of
this redox process on the ligand leads to a much stronger
dependence of the corresponding redox potential on the tpy
+
(
1(PF )) and 0.14 V vs FcH/FcH (2(PF ); see Table 3).
6
6
Additionally, an irreversible oxidation occurs for both
complexes at higher potentials. The first oxidation is ascribed
M
DOI: 10.1021/acs.inorgchem.5b01151
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 8. DFT-calculated spin densities (B3LYP/def2-TZVP/DKH/COSMO(acetonitrile), contour value: 0.01) of 1 (blue, left) and 2 (red, right)
and experimental X-band EPR spectra (ν ≈ 9.4 GHz) obtained from frozen acetonitrile solutions of 1 and 2 (c = 5 mM) generated in situ with
Co(Cp*) . CH hydrogen atoms are omitted for clarity.
2
2
+
2+
Table 4. Electron Paramagnetic Resonance Spectroscopic Parameters of 1 , 2 , 1, and 2 Determined Experimentally (in
a
Frozen Acetonitrile Solution at 77 K) and Theoretically (B3LYP, def2-TZVP, DKH, COSMO(Acetonitrile))
A1,2,3(99,101Ru)/MHz (A /
A1,2,3( N,1)/MHz (A /
14
A1,2,3( N,2)/MHz (A /
14
iso
iso
iso
b
c
d
d
d
g1,2,3
giso
Δg
MHz)
MHz)
MHz)
12+
expt.
2.238, 2.176, 2.045
2.153
2.168
2.345
0.193
84, 140, 112 (112)
88, 162, 91 (114)
94, 187, 103 (128)
28, 36, 73 (46)
41, 41, 63 (48)
8, 8, 8 (8)
calcd. 2.312, 2.172, 2.020
calcd. 2.586, 2.427, 2.021
0.292
22+
1
0.565
expt.
calcd. 2.0147, 2.0001, 1.9414
expt. 2.0008, 1.9998, 1.9685, 1.9897
calcd. 2.0034, 2.0009, 1.9963 2.0002
2.0008, 1.9921, 1.9594, 1.9841
0.0414
0.0733
0.0323
0.0071
17, 10, 45 (24)
e
1.9854
6, 5, 48 (19)
2
3, 7, 31 (14)
3, 6, 39 (16)
f
f
11, 8, 6 (8)
0, 0, 29 (10)
2, 1, 46 (16)
a
14
14
b
For theoretically determined hyperfine coupling constants A( N), only the largest values (A ( N) > 10 MHz) are given. g = (g + g + g )/3.
iso
iso
1
2
3
c
d
e
f
Δg = g − g . A = (A + A + A )/3. Hyperfine coupling to the central pyridine nitrogen atom of the tpy ligand. Hyperfine coupling to one
1
3
iso
1
2
3
peripheral pyridine nitrogen atom of the tpy ligand.
+
functional groups than in the case of the metal-centered
oxidation. Changing the functional group on the tpy ligand
from ethyl carboxy (1(PF )) to N-acetyl amino (2(PF )) shifts
acetonitrile, −1.94 V vs FcH/FcH in dichloromethane). To
study the nature of the oxidized and reduced species EPR
spectra were recorded. Solutions were prepared at a 5 mM
sample concentration with 0.9 equiv of the respective oxidant
or reductant. While both complexes show EPR signals at 77 K
6
6
the first reduction potential by 0.14 V to more negative values
(cf. MO diagrams in Supporting Information Figure S22).
2
+
The oxidation steps in the potential range of −1.0 to −0.5 V
after being reduced to 1 and 2 (see Figure 8), only 1 is
detected via X-band EPR spectroscopy at 77 K (see Supporting
Information Figure S30). At room temperature in dichloro-
methane solution all samples were EPR-silent.
for 1(PF ) following the irreversible reduction processes are
6
similar to those observed for bis(terpyridine)ruthenium(II)
complexes bearing amide functional groups and might be
associated with reduction of the NH proton at the terpyridine
The EPR signal of 12+ (Supporting Information Figure S30;
Table 4) is highly anisotropic (Δg = 0.193) and very broad
indicating a strong contribution of metal d-orbitals to the spin
density. Hyperfine couplings are not resolved in the spectrum,
but a good estimate of the coupling constants of the electronic
88
moiety to hydrogen.
Electron Paramagnetic Resonance Studies on Redox
Products. As evidenced from the cyclic voltammograms, both
2
+
2+
complexes can be oxidized to 1 and 2 using tris(4-
99
101
bromophenyl)aminium hexachloridoantimonate as oxidant
spin to the nuclear spin of ruthenium ( Ru and Ru: I = 5/2,
natural abundance: 30%) is obtained by simulations. This
Ru) = 84, 140, 112 MHz) is large
underlining the strong metal contribution to the radical
+
+
(
0.67 V vs FcH/FcH in acetonitrile, 0.70 V vs FcH/FcH in
9
9,101
dichloromethane) and reduced to 1 and 2 using decamethylco-
coupling (A1,2,3
(
+
baltocene Co(Cp*) as reductant (−1.91 V vs FcH/FcH in
2
N
DOI: 10.1021/acs.inorgchem.5b01151
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
character. Theoretical g tensor and hyperfine coupling
parameter calculations on the DFT wave function generated
at the B3LYP/def2-TZVP/DKH/COSMO(acetonitrile) level
of theory are in excellent agreement with the experimentally
range absorption spectrum is dominated by MLCT transitions
to the electron-poor terpyridine (Ru → tpy) as well as the
electron-rich dipyridylphenyl ligand (Ru → dpb), which was
evidenced by a combined DFT and rR spectroscopic approach.
Theoretical calculations additionally suggest a symmetry-
2
+
determined quantities. 2 is EPR-silent at 77 K and at room
temperature, but theory predicts similar g values and hyperfine
1
forbidden and hence experimentally undetected LL′CT as
2
+
coupling constants to ruthenium as observed for 1 (Table 4).
EPR spectra of 1 and 2 obtained upon reduction of the
respective cations (Figure 8) are substantially sharper and
better-resolved. The g-value anisotropy (Δg = 0.0414 for 1, Δg
lowest spin-allowed optical transition in the red spectral range.
The first oxidation is metal-centered in both complexes 1(PF6)
and 2(PF ) with substantial contribution from the central
6
phenyl ring of the dpb ligand, which corresponds to the ground
state HOMO in both cases. The reduction is tpy-centered with
the unpaired electron localized on the central pyridine ring in
=
0.0323 for 2) is reduced by a factor of 5 compared to the EPR
2+
spectrum of 1 indicating a significantly stronger ligand-based
character of the radical. Hyperfine couplings to the ruthenium
center and one or two nitrogen atoms for 1 and 2, respectively,
are well-resolved and were determined by simulations of the
experimental spectra (see Table 4). The observed spectra are
easily explained with the reduction occurring at the tpy ligand
1(PF ), while it is delocalized over both peripheral pyridine
6
+
rings of tpy in 2(PF ). Reduction of 2 to 2 reverses the order
6
of the unoccupied orbitals LUMO and LUMO+1 as they are
+
close in energy in 2 resulting in a characteristic fingerprint in
the respective EPR spectra.
(
cf. MO diagram in Supporting Information Figure S22). Spin
While both isomers have similar absorption and electro-
chemical characteristics they differ fundamentally in their
density calculations for the neutral complexes 1 and 2 explain
the occurrence of a single nitrogen hyperfine coupling in 1,
while 2 shows couplings to two chemically different nitrogen
atoms. The unpaired electron in 1 is essentially localized at the
central pyridine ring of the tpy−COOEt ligand with its highest
coefficient at the nitrogen p−π orbital leading to strong
anisotropic superhyperfine coupling to that nitrogen nucleus,
excited-state and emission behavior. 1(PF ) is nonemissive
6
both at room temperature and at 77 K, while 2(PF ) shows a
6
very weak emission at room temperature and a much stronger
luminescence at 77 K. Temperature-dependent emission
spectroscopy revealed that a very low activation barrier of ca.
−1
11 kJ mol for the thermal deactivation of the emissive
3
3
which leads to a distinctive triplet splitting of the g signal
MLCT state via a MC state is responsible for the measurable,
3
(Table 4). In 2, however, the unpaired electron is delocalized
but low, emission quantum yield at room temperature.
over the two peripheral pyridine rings of tpy−NHCOMe in 2
with a nodal plane orthogonal to the ligand plane containing
the metal center. Consequently, the EPR signal is much less
well-resolved especially because the superhyperfine coupling
constants to the two peripheral nitrogen nuclei differ (Figure
For 1(PF ) a completely different picture emerges. The
stronger push−pull substitution substantially raises the MC
states in energy, which should lead to an increase in emission
6
3
intensity compared to 2(PF ). Indeed, DFT calculations find
6
3
the MC states high in energy. Hence, these do not contribute
8
). This renders unambiguous determination of all super-
to the efficient excited-state deactivation in 1(PF ). Instead, an
6
3
hyperfine and hyperfine coupling constants (except A of the
unrecognized LL′CT state was found to be lower in energy
3
3
3
two nitrogen atoms) in 2 rather challenging, so these are
estimated from line width and broadening (Table 4). The
electron-donating effect of the N-acetyl amino group attached
to the terpyridine ligand in 2 increases the electron density at
the central pyridine ring and consequently varies the character
than the MLCT state in 1(PF ). The LL′CT state undergoes
6
radiationless deactivation as the radiative relaxation is
symmetry-forbidden (dark state).
3
In essence, cyclometalation using dpb ligands shifts the MC
state above the MLCT for 1 compared to [Ru(tpy) ] thus
3
+
2+
2
and symmetry of the singly occupied molecular orbital
reducing the emission quenching via thermal depopulation
+
3
(
SOMO). It resembles the LUMO+1 of 2 (see Supporting
through MC states. At the same time it generates low-lying
3
Information, Figure S22), whereas the SOMO of 1 is similar to
LL′CT states that evolve radiationless into the ground state
due to the symmetry-forbidden character of the transition
imposed by the orthogonality of the ligands. This quenching via
+
the LUMO of 1 (DFT: B3LYP/def2-TZVP/DKH/COSMO-
(
acetonitrile)). The superhyperfine coupling constants to
3
nitrogen are smaller for 2 than for 1 because of this spin
a
LL′CT state is dominant for push−pull substituted
delocalization over two pyridine rings (spin dilution).
cyclometalated bis(tridentate)ruthenium(II) complexes of
ruthenium, and the underlying mechanisms should be trans-
Furthermore, the Ru−N
distances are larger than the
distances leading to a reduction of the spin−orbit
terminal
26
Ru−N
central
ferable to iridium(III) as well, where similar nonemissive
behavior has been observed.
coupling affecting the unpaired electron and consequently a
lowering of the ruthenium hyperfine coupling in 2 and the g-
value anisotropy. Since the amide bridge is rigid and rotation
Despite the fact that [Ru(dpb)(tpy)]+ complexes are
nonemissive or only weakly emissive at room temperature,
the charge-separation at the Franck−Condon point and the
high reducing potential of the excited state, both induced by the
cyclometalation, render these complexes promising candidates
as sensitizers in photoredox applications.
about the Namide−C
bond is slow at the EPR time scale
especially in rigid matrix) the spin density is asymmetric,
terpyridine
(
which explains the slight differences in coupling constants to
the two peripheral coordinating nitrogen atoms.
CONCLUSION
ASSOCIATED CONTENT
Supporting Information
Experimental procedure for the synthesis of L3; character-
■
■
The key properties of the cyclometalated bis(tridentate)-
ruthenium(II) complexes [Ru(dpb-NHCOMe)(tpy-
*
S
+
+
and [Ru(dpb-COOEt)(tpy-NHCOMe)]+
+
COOEt)]
1
2
ization of 3(PF ); discussion of the synthetic procedures of L1,
6
1
13
were revealed by introduction of electron-donating and
electron-accepting functional groups in the ligand periphery
L2 and L3; FD mass spectra, H and C NMR spectra of L1,
1
13
L2 and L3; ESI mass spectra, H, C, COSY, HSQC, and
+
1
13
of [Ru(tpy)(dpb)] complexes. For both isomers the visible-
HMBC NMR spectra of 1(PF ) and 2(PF ); H and C NMR
6
6
O
DOI: 10.1021/acs.inorgchem.5b01151
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
spectra of 3(PF ); solid-state IR spectra and room-temperature
(19) Meggers, E. Chem. - Eur. J. 2010, 16, 752−758.
20) Wadman, S. H.; Lutz, M.; Tooke, D. M.; Spek, A. L.; Hartl, F.;
Havenith, R. W. A.; van Klink, G. P. M.; van Koten, G. Inorg. Chem.
009, 48, 1887−1900.
21) Heinze, K.; Hempel, K.; Beckmann, M. Eur. J. Inorg. Chem.
006, 2006, 2040−2050.
22) Breivogel, A.; Forster, C.; Heinze, K. Inorg. Chem. 2010, 49,
052−7056.
23) Breivogel, A.; Meister, M.; Fo
Chem. - Eur. J. 2013, 19, 13745−13760.
(24) Abrahamsson, M.; Jager, M.; O
Persson, P.; Becker, H.-C.; Johansson, O.; Hammarstro
Chem. Soc. 2006, 128, 12616−12617.
25) Schramm, F.; Meded, V.; Fliegl, H.; Fink, K.; Fuhr, O.; Qu, Z.;
Klopper, W.; Finn, S.; Keyes, T. E.; Ruben, M. Inorg. Chem. 2009, 48,
677−5684.
26) Wilkinson, A. J.; Puschmann, H.; Howard, J. A. K.; Foster, C. E.;
6
(
emission spectra of 1(PF ) and 2(PF ); decomposition of the
6
6
absorption spectra of 1(PF ) and 2(PF ); MO diagrams of
6
6
2
(
2
(
7
(
1
(PF ) and 2(PF ); TD-DFT excitations of 1(PF ) and 2(PF )
6 6 6 6
with difference density plots; cyclic and square wave
2
+
voltammograms of 1(PF ) and 2(PF ); EPR spectrum of 1 ;
6
6
̈
2
+
3+
3+
DFT spin-density plots of 2 , 2 and 2 ; Cartesian
+
+
2+
2+
2+
3+
coordinates of geometry-optimized 1 , 1 , 2 , 1 , 2 , 1 ,
̈
rster, C.; Laquai, F.; Heinze, K.
3
+
3
3
2
, 1, and 2, as well as of the triplet states LL′CT, MLCT,
3 + 3 3 +
and MC of 1 and MLCT and MC of 2 . The Supporting
Information is available free of charge on the ACS Publications
website at DOI: 10.1021/acs.inorgchem.5b01151.
̈
̈
sterman, T.; Eriksson, L.;
m, L. J. Am.
̈
(
AUTHOR INFORMATION
Corresponding Author
Fax: +49613127277. E-mail: katja.heinze@uni-mainz.de.
■
*
5
(
Williams, J. A. G. Inorg. Chem. 2006, 45, 8685−8699.
(27) Breivogel, A.; Kreitner, C.; Heinze, K. Eur. J. Inorg. Chem. 2014,
2014, 5468−5490.
Author Contributions
The manuscript was written through contributions of all
authors; rR spectra were measured by E.E. and W.W.S. All
authors have given approval to the final version of the
manuscript.
(
2
(
2
28) Bolink, H. J.; Cappelli, L.; Coronado, E.; Gavin
005, 44, 5966−5968.
29) Heinze, K.; Hempel, K.; Breivogel, A. Z. Anorg. Allg. Chem.
009, 635, 2541−2549.
̃
a, P. Inorg. Chem.
Notes
(
30) Englman, R.; Jortner, J. Mol. Phys. 1970, 18, 145−164.
The authors declare no competing financial interest.
(31) Caspar, J. V.; Kober, E. M.; Sullivan, B. P.; Meyer, T. J. J. Am.
Chem. Soc. 1982, 104, 630−632.
ACKNOWLEDGMENTS
(32) Caspar, J. V.; Meyer, T. J. J. Phys. Chem. 1983, 87, 952−957.
■
(
̈
33) Parada, G. A.; Fredin, L. A.; Santoni, M.-P.; Jager, M.; Lomoth,
Parts of this research were conducted using the supercomputer
Mogon and advisory services offered by Johannes Gutenberg
Univ. Mainz (www.hpc.uni-mainz.de), which is a member of
the AHRP and the Gauss Alliance e.V. This work was
financially supported by the Deutsche Forschungsgemeinschaft
R.; Hammarstro
013, 52, 5128−5137.
34) Constable, E. C.; Holmes, J. M. J. Organomet. Chem. 1986, 301,
03−208.
35) Beley, M.; Collin, J. P.; Sauvage, J. P. Inorg. Chem. 1993, 32,
539−4543.
36) King, K. A.; Spellane, P. J.; Watts, R. J. J. Am. Chem. Soc. 1985,
107, 1431−1432.
37) King, K. A.; Watts, R. J. J. Am. Chem. Soc. 1987, 109, 1589−
590.
38) Lamansky, S.; Djurovich, P.; Murphy, D.; Abdel-Razzaq, F.; Lee,
̈
m, L.; Johansson, O.; Persson, P.; Ott, S. Inorg. Chem.
2
(
2
(
4
(
(
GSC 266, Materials Science in Mainz, scholarship for C.K.).
REFERENCES
■
(
1
(
(
1) Brandt, W. W.; Dwyer, F. P.; Gyarfas, E. D. Chem. Rev. 1954, 54,
9
59−1017.
(
(
2) van Houten, J.; Watts, R. J. Inorg. Chem. 1978, 17, 3381−3385.
3) Durham, B.; Caspar, J. V.; Nagle, J. K.; Meyer, T. J. J. Am. Chem.
H.-E.; Adachi, C.; Burrows, P. E.; Forrest, S. R.; Thompson, M. E. J.
Am. Chem. Soc. 2001, 123, 4304−4312.
Soc. 1982, 104, 4803−4810.
(
(
39) Lowry, M. S.; Bernhard, S. Chem. - Eur. J. 2006, 12, 7970−7977.
40) Constable, E. C.; Housecroft, C. E. Polyhedron 1990, 9, 1939−
(
(
5
(
(
(
(
2
(
4) Kalyanasundaram, K. Coord. Chem. Rev. 1982, 46, 159−244.
5) Caspar, J. V.; Meyer, T. J. J. Am. Chem. Soc. 1983, 105, 5583−
1
947.
590.
(
41) Bomben, P. G.; Robson, K. C. D.; Sedach, P. A.; Berlinguette, C.
P. Inorg. Chem. 2009, 48, 9631−9643.
42) Bomben, P. G.; Koivisto, B. D.; Berlinguette, C. P. Inorg. Chem.
010, 49, 4960−4971.
43) Chirdon, D. N.; McCusker, C. E.; Castellano, F. N.; Bernhard,
6) Meyer, T. J. Pure Appl. Chem. 1986, 58, 1193−1206.
7) Kasha, M. Discuss. Faraday Soc. 1950, 9, 14−19.
(
2
(
8) Demas, J. N.; Taylor, D. G. Inorg. Chem. 1979, 18, 3177−3179.
9) Yoon, S.; Kukura, P.; Stuart, C. M.; Mathies, R. A. Mol. Phys.
006, 104, 1275−1282.
10) Lytle, F. E.; Hercules, D. M. J. Am. Chem. Soc. 1969, 91, 253−
57.
11) Suzuki, K.; Kobayashi, A.; Kaneko, S.; Takehira, K.; Yoshihara,
T.; Ishida, H.; Shiina, Y.; Oishi, S.; Tobita, S. Phys. Chem. Chem. Phys.
009, 11, 9850−9860.
12) Lashgari, K.; Kritikos, M.; Norrestam, R.; Norrby, T. Acta
Crystallogr., Sect. C: Cryst. Struct. Commun. 1999, 55, 64−67.
13) Stone, M. L.; Crosby, G. A. Chem. Phys. Lett. 1981, 79, 169−
73.
14) Winkler, J. R.; Netzel, T. L.; Creutz, C.; Sutin, N. J. Am. Chem.
Soc. 1987, 109, 2381−2392.
15) Maestri, M.; Armaroli, N.; Balzani, V.; Constable, E. C.;
Thompson, A. M. W. C. Inorg. Chem. 1995, 34, 2759−2767.
16) Amini, A.; Harriman, A.; Mayeux, A. Phys. Chem. Chem. Phys.
004, 6, 1157−1164.
17) Sun, P.; Krishnan, A.; Yadav, A.; Singh, S.; MacDonnell, F. M.;
Armstrong, D. W. Inorg. Chem. 2007, 46, 10312−10320.
18) Gong, L.; Mulcahy, S. P.; Harms, K.; Meggers, E. J. Am. Chem.
Soc. 2009, 131, 9602−9603.
S. Inorg. Chem. 2013, 52, 8795−8804.
2
(
(44) Colombo, M. G.; Hauser, A.; Guedel, H. U. Inorg. Chem. 1993,
32, 3088−3092.
(45) Hay, P. J. J. Phys. Chem. A 2002, 106, 1634−1641.
(46) Flamigni, L.; Barbieri, A.; Sabatini, C.; Ventura, B.; Barigelletti,
F. Top. Curr. Chem. 2007, 281, 143−203.
2
(
(47) Wadman, S. H.; Kroon, J. M.; Bakker, K.; Lutz, M.; Spek, A. L.;
van Klink, G. P. M.; van Koten, G. Chem. Commun. 2007, 1907−1909.
(48) Wadman, S. H.; Kroon, J. M.; Bakker, K.; Havenith, R. W. A.;
van Klink, G. P. M.; van Koten, G. Organometallics 2010, 29, 1569−
1579.
(
1
(
(
(49) Bruno, N. C.; Tudge, M. T.; Buchwald, S. L. Chem. Sci. 2013, 4,
916−920.
(
2
(
(50) Chavarot, M.; Pikramenou, Z. Tetrahedron Lett. 1999, 40,
6865−6868.
(51) Constable, E. C.; Ward, M. D. J. Chem. Soc., Dalton Trans. 1990,
1405−1409.
(
(52) Fallahpour, R.-A. Eur. J. Inorg. Chem. 1998, 1998, 1205−1207.
(53) Suda, K.; Akagi, K. Macromolecules 2011, 44, 9473−9488.
P
DOI: 10.1021/acs.inorgchem.5b01151
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(
54) Fulmer, G. R.; Miller, A. J. M.; Sherden, N. H.; Gottlieb, H. E.;
Nudelman, A.; Stoltz, B. M.; Bercaw, J. E.; Goldberg, K. I.
Organometallics 2010, 29, 2176−2179.
(
(
(
(
55) Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877−910.
56) Stoll, S.; Schweiger, A. J. Magn. Reson. 2006, 178, 42−55.
57) Neese, F. WIREs Comput. Mol. Sci. 2012, 2, 73−78.
58) Neese, F.; Wennmohs, F.; Hansen, A.; Becker, U. Chem. Phys.
2
(
(
3
(
1
(
2
009, 356, 98−109.
59) Izsak, R.; Neese, F. J. Chem. Phys. 2011, 135, 144105.
60) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77,
865−3868.
61) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1997, 78,
396.
62) Schaf
571−2577.
63) Schaf
829−5835.
64) Andrae, D.; Hau
Theoret. Chim. Acta 1990, 77, 123−141.
65) Weigend, F.; Ahlrichs, R. Phys. Chem. Chem. Phys. 2005, 7,
297−3305.
66) Sinnecker, S.; Rajendran, A.; Klamt, A.; Diedenhofen, M.;
́
̈
er, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97,
(
5
(
̈
er, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994, 100,
̈
ßermann, U.; Dolg, M.; Stoll, H.; Preuß, H.
(
3
(
Neese, F. J. Phys. Chem. A 2006, 110, 2235−2245.
(
(
67) Vlce
̌ ́ ̌
k, A.; Zalis, S. Coord. Chem. Rev. 2007, 251, 258−287.
68) Magyar, R. J.; Tretiak, S. J. Chem. Theory Comput. 2007, 3, 976−
9
(
87.
69) Tao, J.; Perdew, J. P.; Staroverov, V. N.; Scuseria, G. E. Phys.
Rev. Lett. 2003, 91, 146401.
(
(
(
70) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652.
71) Adamo, C.; Barone, V. J. Chem. Phys. 1999, 110, 6158−6170.
72) Yanai, T.; Tew, D. P.; Handy, N. C. Chem. Phys. Lett. 2004, 393,
5
(
1−57.
73) Douglas, M.; Kroll, N. M. Ann. Phys. 1974, 82, 89−155.
74) Hess, B. Phys. Rev. A: At., Mol., Opt. Phys. 1985, 32, 756−763.
75) Hess, B. Phys. Rev. A: At., Mol., Opt. Phys. 1986, 33, 3742−3748.
76) Jansen, G.; Hess, B. Phys. Rev. A: At., Mol., Opt. Phys. 1989, 39,
(
(
(
6
(
016−6017.
77) Pantazis, D. A.; Chen, X.-Y.; Landis, C. R.; Neese, F. J. Chem.
Theory Comput. 2008, 4, 908−919.
(78) Petrenko, T.; Neese, F. J. Chem. Phys. 2007, 127, 164319.
(79) Petrenko, T.; Neese, F. J. Chem. Phys. 2012, 137, 234107.
(80) Beley, M.; Collin, J. P.; Louis, R.; Metz, B.; Sauvage, J. P. J. Am.
Chem. Soc. 1991, 113, 8521−8522.
81) Koivisto, B. D.; Robson, K. C. D.; Berlinguette, C. P. Inorg.
Chem. 2009, 48, 9644−9652.
82) Dreuw, A.; Weisman, J. L.; Head-Gordon, M. J. Chem. Phys.
003, 119, 2943.
(
(
2
(
83) Tozer, D. J. J. Chem. Phys. 2003, 119, 12697−12699.
(
84) Heinze, K.; Hempel, K.; Tschierlei, S.; Schmitt, M.; Popp, J.;
Rau, S. Eur. J. Inorg. Chem. 2009, 2009, 3119−3126.
85) Preiß, J.; Jager, M.; Rau, S.; Dietzek, B.; Popp, J.; Martínez, T.;
Presselt, M. ChemPhysChem 2015, 16, 1395−1404.
86) Egidi, F.; Bloino, J.; Cappelli, C.; Barone, V. J. Chem. Theory
Comput. 2014, 10, 346−363.
87) Heller, E. J.; Sundberg, R.; Tannor, D. J. Phys. Chem. 1982, 86,
822−1833.
88) Kreitner, C.; Grabolle, M.; Resch-Genger, U.; Heinze, K. Inorg.
Chem. 2014, 53, 12947−12961.
89) Breivogel, A.; Hempel, K.; Heinze, K. Inorg. Chim. Acta 2011,
74, 152−162.
(
̈
(
(
1
(
(
3
Q
DOI: 10.1021/acs.inorgchem.5b01151
Inorg. Chem. XXXX, XXX, XXX−XXX