Femtosecond Study of Dimolybdenum Paddlewheel Compounds with Amide/Thioamide Ligands: Symmetry, Electronic Structure, and Charge Distribution in the1MLCT S1 State

Article abstract of DOI:10.1021/acs.inorgchem.6b02517

Four photophysically interesting dimolybdenum paddlewheel compounds are synthesized and characterized: I and II contain amide ligand (N,3-diphenyl-2-propynamide), and III and IV contain thioamide ligand (N,3-diphenyl-2-propynethioamide). I and III are tra

Full text of DOI:10.1021/acs.inorgchem.6b02517

Article
pubs.acs.org/IC
Femtosecond Study of Dimolybdenum Paddlewheel Compounds
with Amide/Thioamide Ligands: Symmetry, Electronic Structure, and
1
Charge Distribution in the MLCT S1 State
Changcheng Jiang,* Philip J. Young,† Samantha Brown-Xu,‡ Judith C. Gallucci,
and Malcolm H. Chisholm
Department of Chemistry and Biochemistry, The Ohio State University, 100 West 18th Avenue, Columbus, Ohio 43210, United
States
S
* Supporting Information
ABSTRACT: Four photophysically interesting dimolybde-
num paddlewheel compounds are synthesized and charac-
terized: I and II contain amide ligand (N,3-diphenyl-2-
propynamide), and III and IV contain thioamide ligand
(N,3-diphenyl-2-propynethioamide). I and III are trans-
Mo2L2(O2C-TiPB)2-type compounds, and II and IV are
Mo2L4-type compounds, where O2C-TiPB is 2,4,6-triisopro-
pylbenzoate. I−IV display strong light absorption due to metal
to ligand charge transfer (MLCT) transitions from molybde-
num to the amide/thioamide ligands. Charge transfer
dynamics in the MLCT excited states of I−IV have been
examined using femtosecond transient absorption (fs-TA)
spectroscopy and femtosecond time-resolved infrared (fs-
TRIR) spectroscopy. The asymmetric amide/thioamide ligands show two forms of regioarrangements in the paddlewheel
compounds. Analyses of the ν(CC) bands in the fs-TRIR spectra of I and II show similar electron density distribution over
1
ligands in their MLCT S1 states where only two amide ligands are involved and the transferred electron is mainly localized on
one of them. The fs-TRIR spectra of III and IV, however, show different charge distribution patterns where the transferred
electron is fully delocalized over two thioamide ligands in III and partially delocalized in IV. Fast interligand electron transfer
(ILET) was recognized as the explanation for the various charge distribution patterns, and ILET was shown to be influenced by
both the ligands and the ligand arrangements.
MLCT states (especially the lowest singlet excited states, i.e.,
1MLCT S1 states) due to their simple structure and high
symmetry.14 Unlike many mononuclear complexes, Mo2 and
W2 compounds have 1MLCT S1 states that are sufficiently long
lived to be carefully examined by femtosecond transient
absorption (fs-TA) and femtosecond time-resolved infrared
(fs-TRIR) experiments.15−17 A fair number of Mo2 and W2
compounds with carboxylate ligands have been studied,
typically in the form of trans-M2L2L′2 and M2L4 (with two
and four acceptor ligands, respectively, as shown in Figure 1). A
bulky carboxylate ligand with twisted accepting groups, 2,4,6-
triisopropylbenzoate (O2C-TiPB), was frequently used as an
auxiliary ligand in the trans-M2L2L′2 compounds for synthetic
convenience and also because the ligand only allows high-
energy MLCT transitions.18
INTRODUCTION
■
Transition metal (TM) compounds showing strong light
absorption due to MLCT transitions have found wide
applications as photosensitizers, photocatalysts, and lumines-
cent materials.1−4 Upon absorption of a photon at the MLCT
bands, an electron moves from a metal-based orbital into a
ligand-based orbital which leaves the compounds in charge-
separated MLCT states. With multiple acceptor ligands, the
transferred electron in the MLCT states of TM compounds can
have various distribution patterns: the electron could be
localized on one ligand or delocalized over two or more
ligands.5−9 Such differences would strongly influence the
photophysical and photochemical properties of TM com-
pounds, yet charge distributions in MLCT states have not been
routinely addressed due to quick relaxation of the excited states.
Similar localization−delocalization problems also have been
observed in other systems such as the excited states of branched
organic compounds and also the ground states of certain
radicals.10−13
With IR reporter groups such as CC and CN bonds,
additional electron density on the ligands in the excited states
of trans-M2L2(O2C-TiPB)2 and M2L4 compounds can be
Mo2 and W2 quadruply bonded compounds provide
convenient models for the study of charge dynamics in their
Received: October 19, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acs.inorgchem.6b02517
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 2. Structure of N,3-diphenyl-2-propynamide, N,3-diphenyl-2-
propynethioamide, and 2,4,6-triisopropylbenzoate ligands.
compounds II and IV,26 it was found that the amide/thioamide
ligands are centrosymmetrically disposed to each other in I−III
while adopting a mirror-symmetric arrangement in IV. The
photophysical properties of I−IV were thoroughly examined
using steady-state absorption, emission, and time-resolved
spectroscopic methods. The excited-state charge distribution
patterns were found to be similar in I and II but distinctively
different in III and IV. Electronic transitions of I−IV were
investigated with density functional theory (DFT) and time-
dependent density functional theory (TD-DFT) methods.
Correlations between electronic structures and excited-state
charge dynamics are discussed.
Figure 1. General structure of trans-M2L2(O2C-TiPB)2 and M2L4
paddlewheel compounds with carboxylate ligands where M is Mo or
W.
directly observed through the shift of ν(CC) and ν(CN)
bands in fs-TRIR spectra.16,19,20 Both localized and delocalized
1MLCT S1 states were found in these compounds with different
acceptor ligands.14,21 The transition from localized charge
distribution to delocalized charge distribution appears to be
connected with interligand electron transfer (ILET) on the IR
time scale (∼10−12 s) where “fast” ILET would present as if the
electron were delocalized while “slow” ILET would present as if
the electron were localized.5,22 Such ligand-based electron
transfer process is supported by the presence of mixed-valence
(MV) states in Mo2 and W2 paddlewheel radical anions.23 The
bridging units, which include the dimetal center and ligand
donor atoms, were determined to play an important role in the
ILET process.21 Also, the fs-TRIR spectra of trans-M2L2(O2C-
TiPB)2 and M2L4 compounds with the same acceptor ligands
were generally found to be similar.16,20 Such observations
suggest that only trans ligands are involved in the fast ILET
process.
EXPERIMENTAL SECTION
■
General Procedures. Compounds I−IV are air and moisture
sensitive, especially so in solution. Syntheses and handling of all four
compounds were conducted using standard Schlenk line and glovebox
techniques. All solvents used were thoroughly dried and degassed
following standard protocols.27
Syntheses. The preparative methods for compounds I−IV are
displayed in Scheme 1. Two dimolybdenum carboxylates, Mo2(O2C-
Scheme 1. Preparation of I−IV
Previously, in the amidinate−carboxylate compound trans-
Mo2[(iPrN)2CCCPh]2(O2CMe)2, we discovered that the
1
charge distribution on the amidinate ligands in the MLCT S1
state is a borderline case,16 i.e., the majority of the electron
density is localized on one of the amidinate ligands but some
electron density is also present on the other ligand. Due to the
similarities between this type of excited-state species and a class
II MV compound, we classified it as a class II 1MLCT S1 state.
Follow-up studies on compounds trans-Mo2[(PhN)2CC
CPh]2(O2C-TiPB)2 and Mo2[(PhN)2CCCPh]4 have shown
that the charge distribution can be tuned, and the total
localization case was found in the latter compound.24 These
observations prompted us to investigate compounds with
similar ligand sets.
To understand the role of the bridging units in interligand
electron transfer, we set to change the electronic structure of
the Mo2 compounds. Here, we prepared Mo2 paddlewheel
compounds I−IV with amide and thioamide ligands where one
of the nitrogen atoms in the amidinate ligands is substituted by
an oxygen or a sulfur atom. The amide and thioamide ligands
are shown in Figure 2 together with O2C-TiPB. The 2p orbitals
of oxygen are lower in energy in comparison to nitrogen, and
sulfur instead has more diffuse 3p orbitals. The substitution
created significant electronic structure changes. Since amide
and thioamide ligands are asymmetric compared to carboxylate
and amidinate ligands, different regioarrangements also
appeared in these paddlewheel compounds.25
TiPB)4 and Mo2(OAc)4, were used as starting materials, and they were
synthesized conveniently following literature methods.18,25 I and III
were prepared by reacting Mo2(O2C-TiPB)4 with two equivalents of
ligand (lithium N,3-diphenyl-2-propynamide and lithium N,3-
diphenyl-2-propynethioamide, respectively); II and IV were prepared
by reacting Mo2(OAc)4 with four equivalents of ligand.26 Compounds
I−IV were purified through solution processes, and the purity of each
compound was confirmed by 1H NMR and MALDI-TOF mass
spectra. Details of ligand and compound syntheses and NMR and
MALDI-TOF mass spectra for I−IV are included in the Supporting
Information.
X-ray Crystallography. Crystals of I and III suitable for single-
crystal X-ray diffraction were grown by diffusion of hexanes into
concentrated THF solutions of I and III containing a small amount of
DMSO. Diffraction data were collected on a Nonius Kappa Apex II
CCD diffractometer with Mo Kα radiation. All work was done at 150
K using an Oxford Cryosystems Cryostream Cooler. X-ray crystallo-
graphic data in CIF format can be found in the Supporting
Information. Crystal data and selected bond lengths are also included
as Tables S1 and S2 in the Supporting Information. The structures of
II and IV have been reported previously.26
Absorption and Emission Spectra. Absorption spectra of I−IV
were record on a PerkinElmer Lambda 900 spectrometer. Data were
collected from 250 to 800 nm in THF solution using a 1 cm × 1 cm
quartz cell. The cell is equipped with a Kontes stopcock. Molar
extinction coefficients (ε) of I−IV were calculated from absorbance
The structures of the trans-M2L2L′2 compounds I and III
were studied using single-crystal X-ray crystallography; together
with the previously reported structures of M2L4 homoleptic
B
DOI: 10.1021/acs.inorgchem.6b02517
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
values at four different concentrations. Emission spectra of I−IV were
obtained with the Jabin Yvon Fluorolog 3 fluorometer equipped with a
Synapse CCD detector. Data were collected from 500 to 1100 nm in
2-methyl tetrahydrofuran (2-MeTHF) solution in a J. Young tube at
both room temperature and 77 K with a customized Dewar cuvette.
Electronic Structure Computation. Structure optimizations of
I−IV were performed using density functional theory (DFT) methods
with the B3LYP functional in the Gaussian 09 suite.28 The 6-31G(d)
basis set was used for all nonmetal elements; the SDD energy-
consistent pseudopotentials and the SDD energy-consistent basis set
were used for Mo atoms.29 Atomic coordinates from crystal structures
were used as the initial input. Time-dependent density functional
theory (TD-DFT) calculations were carried out for I−IV to study the
electronic transitions.
Figure 3. Crystal structures of I and II. Green = molybdenum, blue =
nitrogen, red = oxygen, and gray = carbon. Hydrogen atoms and
solvent molecules were obscured for clarity. (Plot of II was adapted
from ref 26 with permission of The Royal Society of Chemistry.)
DFT calculations were also performed on I−IV in their triplet states
and I−IV bearing one negative charge (“molecular anions”, mimics of
the MLCT singlet states) to aid the interpretation of fs-TRIR spectra.
To cut computational costs in these open-shell calculations, phenyl
groups on the nitrogen atoms and aryl groups in the O2C-TiPB ligands
were replaced with hydrogen atoms, and these results will be referred
3
to with a prime notation (for example, I′ represents the calculated
triplet state of I, and I′− represents the calculated molecular anion of
I). All vibrational frequency results were scaled by a factor of 0.961 as
suggested.30
Femtosecond and Nanosecond Time-Resolved Spectrosco-
py. Femtosecond transient absorption spectroscopy (fs-TA) and
femtosecond time-resolved infrared spectroscopy (fs-TRIR) experi-
ments were performed with a Ti:sapphire laser system (1 kHz, fwhm
≈ 300 fs) as described previously.31,32 In the fs-TA experiments,
sample solutions were prepared in THF with an absorbance of ∼0.4 at
the maximum of the MLCT band and kept in a 1.0 mm quartz cuvette
with a Kontes stopcock. In the fs-TRIR experiments, the sample
solutions were prepared in THF as well with an absorbance of ∼1 at
the maximum of the MLCT band and were sealed in a PerkinElmer
semidemountable cell with a 0.1 mm Teflon spacer between two 4 mm
CaF2 windows. The excitation wavelength was set at 515 nm for I−III
and at 568 nm for IV in both fs-TA and fs-TRIR experiments, and the
excitation power was set at 1−2 μJ at the samples. fs-TA spectra from
300 to 700 nm and fs-TRIR spectra from 1350 to 2250 cm−1 were
collected for I−IV.
Nanosecond transient absorption (ns-TA) spectra were performed
on a home-built instrument equipped with a Spectra-Physics GCR-150
Nd:YAG laser source (fwhm ≈ 10 ns) and a Hamamatsu R928
photomultiplier tube. Transient signals were processed with a
Tektronics 400 MHz oscilloscope. The excitation wavelengths were
set at 532 nm for I−IV, and the excitation power was set at ∼100 mW
at the samples. Sample solutions were prepared in THF with an
absorbance of ∼0.3 at the maximum of the MLCT band and kept in a
1 cm × 1 cm quartz cuvette with a Kontes stopcock. ns-TA spectra
from 380 to 760 nm were collected for I−IV.
Figure 4. Crystal structures of III and IV. Green = molybdenum, blue
= nitrogen, red = oxygen, gray = carbon, and yellow = sulfur.
Hydrogen atoms and solvent molecules were obscured for clarity.
(Plot of IV was adapted from ref 26 with permission of The Royal
Society of Chemistry.)
symmetry, respectively. The two thioamide ligands in III are
centrosymmetrically disposed to each other, similar to I. Each
set of the trans-thioamide ligands in IV instead has a mirror-
reflection relationship. Overall IV adopts the “trans-2,2”
structure.26
As shown by Berry and co-workers, the configuration of Mo2
amide/thioamide compounds changes with different amide and
thioamide ligands.33 However, no isomers were isolated in the
synthetic process, and no isomers are expected to exist in
solution as a result of the nonlabile nature of the Mo−N
bonds.26,33
In all four compounds, extended conjugation is expected
along the trans-L−Mo2−L chains. The metal atoms and the N−
C−X bridging ligands are almost coplanar with the CC
Ph moieties. This is more pronounced in I and III than in II
and IV. The phenyl groups on the nitrogen atoms are twisted
by ∼60° from the N−C−X planes in all four compounds,
effectively removing them from contributing to the conjugation.
This is also true for the aryl groups in the O2C-TiPB ligands,
which are twisted by 80−90° from the CO2 planes.
Compounds I−IV show Mo−Mo bond lengths of ∼2.10 Å,
which is typical of Mo2 quadruple bonds.25 Average Mo−N
bond length in all compounds is ∼2.13 Å; average Mo−O bond
length is ∼2.10 Å (between Mo2 and amide ligands), and
average Mo−S bond length is ∼2.46 Å. Due to the longer Mo−
S bonds, the CCPh groups in III and IV deviate slightly
from the perpendicular position with respect to the Mo−Mo
quadruple bond. Distance between the centers of CC bonds
on trans amide/thioamide ligands is ∼9.6 Å in I and II and
∼10.0 Å in III and IV.
All time-resolved spectra were plotted in Igor Pro 6.0. Kinetic traces
were fitted with a global fitting package with a sum of exponentials,
S(t) = ∑iAi exp(−1/τi) + C, where Ai is the amplitude, τi is the
lifetime, and C is an offset.
RESULTS AND DISCUSSION
■
X-ray Crystal Structure. With the asymmetric bridging
ligands, the core structures of I, II, III, and IV show reduced
symmetry from the ideal D2h and D4h symmetry for trans-
M2L2L′2 and M2L4 compounds.
The crystal structures of I and II are shown in Figure 3. The
core structures of I and II have the same C2h pseudo symmetry.
The two amide ligands in I are centrosymmetrically disposed to
each other, and the same arrangement is found for each set of
the trans-amide ligands in II. Overall, II adopts the so-called
“cis-2,2” structure.26
The crystal structures of III and IV are shown in Figure 4.
The core structures of III and IV have the C2h and D2d pseudo
C
DOI: 10.1021/acs.inorgchem.6b02517
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 5. MO energy diagrams and selected orbital surfaces of I−IV. MO surfaces were plotted with Gaussian 5.08, and idealized symmetries were
assigned. From top to bottom, surface plots of LUMO+2 (au, δ*), LUMO+1 (bg, LLπ*), LUMO (au, LLπ*), and HOMO (bg, δ) are shown for I and
III; LUMO+4 (au, δ*), LUMO+1 (au, LLπ*), LUMO (bu, LLπ*), and HOMO (bg, δ) are shown for II; LUMO+2 (e, LLπ*), LUMO+1 (e, LLπ*),
LUMO (a2, LLπ*+δ*), and HOMO (b, δ) are shown for IV.
Electronic Structure and Electronic Transitions.
Electronic structures of I−IV were investigated using DFT
methods. Molecular orbital (MO) energy levels, symmetry, and
frontier orbital surface plots are presented in Figure 5. Natural
bonding orbital (NBO) analyses were carried out to study the
orbital compositions.34,35 The contributions of metal orbitals to
the frontier orbitals of I−IV are summarized in Table 1. The
increase in metal−ligand orbital mixing is significant with the
asymmetric ligands and with the introduction of sulfur atoms.
In I and III, the π* orbitals of the two amide/thioamide
ligands combine and transform as au and bg symmetry in the
C2h point group. The Mo2δ orbital transforms as bg symmetry,
and the Mo2δ* orbital transforms as au symmetry. The bg LLπ*
orbital interacts with the Mo2δ orbital and is raised in energy
relative to the au LLπ* orbital. The resulting HOMO is a Mo2δ-
based orbital with bg LLπ* contribution. The resulting LUMO
is essentially the au LLπ* orbital with some Mo2δ* orbital
character.
The increase of metal orbital character in the LUMO of I and
III is significant with the symmetry reduction.20,24 Also, due to
the lower π* orbital energy of the thioamide ligand and the
more diffuse p orbitals of the sulfur atom, III shows a lower
energy HOMO, a lower energy LUMO, and stronger metal−
ligand orbital mixing in comparison to I. The NBO analysis
shows that the ligand orbitals contribute ∼40% to the HOMO
of III compared to ∼35% to the HOMO of I. In addition, metal
orbitals contribute 9.9% to the LUMO of III compared to 3.4%
to the LUMO of I.
Table 1. Mo2 Contribution (%) to Frontier Molecular
Orbitals Calculated from NBO Analysis
HOMO LUMO LUMO+1 LUMO+2 LUMO+3 LUMO+4
I
65.4
58.6
60.0
51.2
3.4
0.7
7.2
3.3
7.6
1.2
76.2
0.1
II
74.4
39.2
10.1
13.0
III
IV
9.9
69.2
1.2
33.2
In II and IV, the four ligand π* orbitals combine and
transform into four different LLπ* combinations, namely, the
bu, au, bg, and ag LLπ* orbitals in II (C2h point group) and a2, e,
and b1 LLπ* orbitals in IV (D2d point group). The arrangement
and composition of LLπ* orbitals in II and IV appear to be
very different from each other due to the ligand arrangements.
In II, the au and bu LLπ* orbitals become the LUMO and
LUMO+1, and their energy levels are very close. Transitions
from HOMO to LUMO and LUMO+1 are both fully allowed.
Electronic structures of trans-Mo2L2(O2C-TiPB)2 and Mo2L4
are best described by interactions between the ligand π*
orbitals (the combination of which will be referred to as LLπ*
orbitals) and Mo−Mo quadruple bond orbitals.23 The highest
occupied molecular orbital (HOMO) in trans-Mo2L2(O2C-
TiPB)2 and Mo2L4 compounds is typically composed of the
Mo2δ orbital, and the lowest unoccupied molecular orbital
(LUMO) is usually composed of one of the LLπ* orbitals.
a
Table 2. Electronic Transitions in Compounds I−IV Predicted by TD-DFT
origin
λ (nm)
f
I
S0 → S1
S0 → S2
S0 → S10
S0 → S1
S0 → S2
S0 → S1
S0 → S2
S0 → S1
S0 → S2
HOMO → 0.54 LUMO, 0.42 LUMO+2 (δ*)
HOMO → 0.53 LUMO+2 (δ*), 0.44 LUMO
HOMO → 0.57 LUMO+6, 0.41 LUMO+4
HOMO → 0.68 LUMO+1, 0.15 LUMO+4 (δ*)
HOMO → 0.70 LUMO
512
492
352
516
512
570
489
560
521
0.360
0.367
0.285
0.593
0.749
0.421
0.338
0.015
0.556
II
III
IV
HOMO → 0.66 LUMO, 0.22 LUMO+2 (δ*)
HOMO → 0.64 LUMO+2 (δ*), 0.23 LUMO
HOMO → 0.67 LUMO (δ*), 0.16 LUMO+4
HOMO → 0.68 LUMO+1, 0.13 LUMO+2
a
Notes: λ is the transition wavelength, and f is the oscillator strength.
D
DOI: 10.1021/acs.inorgchem.6b02517
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
In IV, the a2 LLπ* orbital with strong Mo2δ* orbital character
turns out to be the LUMO instead of the e LLπ* orbitals
normally found in Mo2 homoleptic carboxylate compounds.14
The Mo2δ* orbital contributes greatly to the LUMO of IV
(33%). The HOMO → LUMO transition in IV is symmetry
allowed with the absence of an inversion center.
TD-DFT calculations were also carried out for compounds
I−IV to study the electronic transitions. The results are
summarized in Table 2.
As expected, the HOMO → LUMO transitions, which are
MLCT in nature, are the dominant transitions in all of the
compounds except IV. In IV, the oscillator strength of the
HOMO → LUMO transition is weak while the HOMO →
LUMO+1 transition appears to be the dominant transition.
Since the LUMO+1/LUMO+2 of IV are similar to the e-type
LUMO/LUMO+1 normally found for homoleptic carboxylate
compounds, the HOMO → LUMO+1 transition in IV is still a
typical MLCT transition. It is worth noting that in the MLCT
excited states of II and IV, Jahn−Teller distortions are expected
to appear when the electron is transferred into the close-to-
degenerate LLπ* orbitals, as described previously.20,24
The calculations also predicted that there is a specific and
intense metal-based δ → δ* transition in I and III at slightly
higher energies of the MLCT transition. The mixing of MLCT
transitions in the metal-based δ → δ* transition is unexpectedly
strong. Generally, the δ → δ* absorption bands were weak in
the previously studied carboxylate and amidinate com-
pounds.20,24,36
III is larger than that of I, agreeing with TD-DFT calculations.
Also in III, the expected MLCT transition to O2C-TiPB ligand
is obscured in the ligand-based transitions. In IV, the strongest
absorption band in the visible region appears at 500 nm (ε =
42 300 M−1 cm−1). The predicted low-intensity HOMO →
LUMO transition at 560 nm indeed cannot be identified. The
transitions in III and IV are generally stronger than those in I
and II, as shown by the molar extinction coefficients.
Emission spectra of I−IV from the visible to the near-
infrared region were taken in 2-MeTHF solution at room
temperature and at 77 K. The spectra are shown in Figure 7,
UV−vis Absorption and Emission. The absorption
spectra of I−IV recorded in THF solution at room temperature
are shown in Figure 6. Generally, the compounds show strong
ligand-based absorption bands in the UV region and MLCT
absorption bands the visible region.
Figure 7. Normalized emission spectra of I−IV. Emission spectra at
room temperature are shown in black, and emission at 77 K are shown
in red. Spectra were recorded in 2-MeTHF.
and the peak wavelengths are summarized in Table 3.
Emissions from both singlet states (fluorescence) and from
triplet states (phosphorescence) were observed for all four
compounds.
Table 3. Emission Peaks of I−IV Recorded in 2-MeTHF at
Room Temperature (RT) and at 77 K
fluorescence (nm)
phosphorescence (nm)
RT
77 K
RT
77 K
I
562
560
604
565
550
551
593
563
952
909
969
912
951
908
946
898
II
III
IV
Figure 6. Normalized UV−vis absorption spectra of I−IV. Spectra
were recorded in THF solution at room temperature.
The fluorescence of I−IV was determined to be from the
3
1MLCT states and the phosphorescence from the δδ* states
Compounds I and II, with amide ligands, show MLCT
absorption bands at 490 and 486 nm, respectively. The molar
extinction coefficient ε was determined to be 18 600 M−1 cm−1
at the MLCT band of I and 31 600 M−1 cm−1 of II. These
MLCT bands have steeper slopes on the low-energy side.
There is a shoulder at the MLCT band of I appearing at ∼440
nm, which was assigned to the δ → δ* transition. The
absorption band at ∼350 nm in I was assigned to the higher
energy MLCT transition to the O2C-TiPB ligands.
Compound III shows two absorption bands in the visible
region with one at 470 nm (ε = 5600 M−1 cm−1) and the other
at 559 nm (ε = 19 700 M−1 cm−1). The 559 nm band was
assigned to the MLCT transition and the 470 nm band to the δ
→ δ* transition. The splitting between the two transitions in
based on their energy, band shape, and vibronic progressions at
low temperature.14 Typically, the fluorescence maxima were
observed 1000−2000 cm−1 lower in energy from the lowest
1MLCT absorption maxima at room temperature. The
phosphorescence maxima are observed around 960 nm for I
and III and around 900 nm for II and IV and are not related to
the MLCT absorption bands. It is interesting that the
phosphorescence (3δδ* state) energies are similar between
the amide and the thioamide compounds despite the
differences between oxygen and sulfur.
Upon cooling in liquid nitrogen, the fluorescence bands of
I−IV shift to higher energy by ∼10 nm while the
phosphorescence bands shift little. Both the singlet and the
E
DOI: 10.1021/acs.inorgchem.6b02517
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 8. fs-TA spectra of I−IV, collected in THF solution at room temperature. λex = 515 nm for I, II, and IV; λex = 568 nm for III.
Figure 9. Kinetic traces of the early absorption feature for I, II, III, and IV. Data are taken at 420, 420, 380, and 439 nm, respectively.
triplet emission bands gain intensity where the singlet emission
is more strongly affected. Vibronic progressions start to reveal
themselves where shoulders of the 1MLCT singlet state
interference. All four compounds show early features decaying
on the picosecond time scale and long-lived features persisting
into the nanosecond time scale, which were attributed to the
lowest singlet state (S1 states) and the lowest triplet states (T1
states), respectively.
3
emission become clearer and sharp peaks in the δδ* triplet
state emission start to emerge. Vibronic coupling modes at
3
∼300 cm−1 were observed in the δδ* state emission in I and
The short time-delay spectra of I−IV in Figure 8 (blue and
purple traces) show distinctive S1 state absorption bands. In I
and II, the S1 state absorption maxima can be determined to be
between 400 and 500 nm, while in III and IV, the absorption
bands are quite broad, making the specific absorption maxima
difficult to identify. A strong negative band can be found at
∼500 nm in I, II, and IV, which was attributed to ground-state
bleach. In III, two negative bands, one at ∼460 nm and one at
∼560 nm, were also attributed to ground-state bleach, both
consistent with the UV−vis spectra. A negative band around
580 nm in I and 600 nm in II were attributed to stimulated
emission from S1 states. The absorption bands and stimulated
emission bands decay within picoseconds in a parallel manner.
The S1 state absorption features of I−IV were all fitted with
monoexponential decay functions. The lifetimes were found to
III. These vibronic features are assigned to totally symmetric
vibrational modes such as Mo−Mo stretches and coordinated
25
3
N−C−X stretches in the δδ* states. In compounds II and
IV, the vibronic features appear to be more complex.
Transient Absorption Spectra and Excited-State Life-
times. The excited states of I−IV were studied in THF
solution by pump−probe-type transient absorption spectrosco-
py in both the femtosecond scale (fs-TA) and the nanosecond
scale (ns-TA).
The fs-TA spectra of I−IV are shown in Figure 8 with time
delays shown inset. Excitation wavelengths were set to 515 nm
for I, II, and IV and 568 nm for III, all to the lower energy side
of their MLCT bands. Data points around the excitation
wavelength in each case were excluded due to pump laser
F
DOI: 10.1021/acs.inorgchem.6b02517
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
1MLCT S1 states. The shift magnitudes of the IR bands on the
ligands are directly correlated with the transferred electron
density, and thus, they were probed carefully.
be 5.7, 6.9, 3.5, and 1.9 ps for I, II, III, and IV, agreeing well
with the S1 lifetimes found for previous related compounds.14
With the spectra features and lifetimes observed, the S1 states of
1
I−IV are all reaffirmed as MLCT S1 states, Figure 9.
From the FT-IR spectra, compounds I−IV show three main
In the longer time-delay spectra shown in Figure 8 (red
traces), the features were attributed to the T1 state absorption
and ground-state bleach. Strong absorption bands can be
identified at ∼350−450 nm and bleach bands at the MLCT
absorption maxima of each compound. The T1 states of I−IV
at longer time scales were studied with ns-TA spectra.
types of IR reporter bands: (1) CC stretches, ν(CC), (2)
phenyl ring stretches (in the CCPh moieties), ν(C6)ring
,
and (3) carboxylate, amide, and thioamide stretches, ν(O−C−
O), ν(N−C−O), and ν(N−C−S). In the ground states of I−
IV, ν(CC) appears as a weak absorption band at ∼2210
cm−1, ν(C6)ring as a moderate absorption band at ∼1600 cm−1,
and ν(O−C−O), ν(N−C−O), and ν(N−C−S) as strong
absorption bands in the 1200−1500 cm−1 region where the
asymmetric modes νas(O−C−O), νas(N−C−O), and νas(N−
C−S) appear at higher energy than the symmetric modes
νs(O−C−O), νs(N−C−O), and νs(N−C−S).
In the ns-TA experiments, an excitation wavelength of 532
nm was used for all compounds. The spectra are shown in
Figures S14−S17 in the Supporting Information. The T1 state
absorption bands and the ground-state bleach bands observed
in ns-TA spectra were at the same positions as the fs-TA
spectra. T1 lifetimes of compounds I−IV were extracted from
the absorption features and found to be 94, 84, 71, and 85 μs,
For the purpose of discussing the features, the fs-TRIR
spectra were divided into two regions: the 1850−2250 cm−1
region containing the ν(CC) bands and the 1350−1650
cm−1 region containing the ν(C6)ring, ν(O−C−O), ν(N−C−
O), and ν(N−C−S) bands.
3
respectively. These lifetimes are consistent with the δδ* T1
assignments, and they are longer than those found in
carboxylate compounds.14
The S1 and T1 lifetimes of I−IV obtained from fs-TA and ns-
TA spectra are shown in Table 4 together with results from the
fs-TRIR spectra (see the Time-Resolved Infrared Spectroscopy
section). The S1 lifetimes from fs-TA and fs-TRIR experiments
agree well.
fs-TRIR Spectra in the ν(CC) Region. fs-TRIR spectra
of I−IV from 1850 to 2250 cm−1 are shown in Figure 10, where
the inverted ground-state FT-IR spectra were also plotted
together as dotted lines. At the picosecond time delays, strong
ν(CC) bands corresponding to the 1MLCT S1 states of I−IV
appear with large shifts to lower energy from their ground-state
frequencies. The peak positions are marked by black arrows.
In compound I, two ν(CC) bands appeared in its 1MLCT
S1 state: a weaker band at around 2152 cm−1 and a stronger
band around 1970 cm−1. Compared to the ground state, the
two ν(CC) bands are shifted to lower energy by 53 and 235
cm−1, respectively. In compound II, the spectra are quite similar
to that of I where also two ν(CC) bands were identified, at
1970 and 2152 cm−1. The shapes of the ν(CC) bands,
however, are noticeably flatter than those of I.
Table 4. Excited-State Lifetimes of I−IV
S1 lifetime (ps)
fs-TRIR
T1 lifetime (μs)
fs-TA
ns-TA
I
5.5 0.2
6.9 0.2
3.5 0.2
1.6 0.1
5.7 0.2
6.9 0.2
3.5 0.2
1.9 0.2
94
84
71
83
5
3
8
2
II
III
IV
In compound III, only one ν(CC) band was observed in
the 1MLCT S1 state. The peak was found at ∼2021 cm−1 and is
noticeably sharp. This ν(CC) band in III is shifted by 189
cm−1 from the ground state. In compound IV, two ν(CC)
Time-Resolved Infrared Spectroscopy. The excited
states of I−IV were also studied with fs-TRIR spectra in
THF solution. Here, special interest is taken on the electron
density distribution over the amide/thioamide ligands in the
Figure 10. fs-TRIR spectra of I−IV in the region of 1850−2250 cm−1. Time delays are shown as inset. Inverted ground-state IR absorption spectra
(gs-IR) are also plotted as dotted lines.
G
DOI: 10.1021/acs.inorgchem.6b02517
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 11. fs-TRIR of I−IV in the region of 1350−1650 cm−1 containing ν(C6)ring, ν(O−C−O), ν(N−C−O), and ν(N−C−S) bands. Inverted
ground-state IR absorption spectra (gs-IR) are also plotted, as dotted lines. Data points around 1450 cm−1 were obscured due to strong solvent
absorption.
Figure 12. Kinetic traces of ν(CC) bands in fs-TRIR spectra of I, II, III, and IV. Data are taken at 1970, 1970, 2021, and 2048 cm−1, respectively.
bands were found, at 2048 and 1974 cm−1. The former is
shifted by 160 cm−1 from the ground state, and the latter is
shifted by 234 cm−1. Again, the two ν(CC) bands in this
homoleptic compound appear to be broader than that of III.
In a picosecond time window, the ν(CC) bands
corresponding to the 1MLCT S1 states of I−IV decay
completely and long-lived weak ν(CC) bands corresponding
S1 states and the ν(O−C−O), ν(N−C−O), and ν(N−C−S)
3
bands in the δδ* T1 states.
The ν(C6)ring bands generally show similar features as ν(C
C) bands. In the 1MLCT S1 states of I−IV, a weak bleach band
at the ground-state ν(C6)ring frequency (∼1595 cm−1) and a
strong absorption band on the lower energy side were found for
ν(C6)ring modes. The shift of ν(C6)ring absorption bands is ∼40
cm−1 in I and II, 19 cm−1 in III, and 10 cm−1 in IV. Because the
phenyl ring is more remote from the bimetal center, ν(C6)ring
3
to the δδ* T1 states emerge at ∼2200 cm−1, close to their
ground-state frequencies. This transformation agrees with the
1MLCT S1 to 3δδ* T1 conversion and the return of the
transferred electron from the ligands to the dimetal center. It
3
features were not observed in the δδ* states.
Over time, the ν(C6)ring features decay and ν(O−C−O)/
3
appears that the ν(CC) absorption bands in the δδ* T1
ν(N−C−O)/ν(N−C−S) features grow in strength. The
1
states are much weaker in comparison to those in the MLCT
1
3
transition is, again, consistent with the MLCT S1 to δδ* T1
conversion. Vibrations corresponding to ν(O−C−O)/ν(N−
C−O)/ν(N−C−S) in I−IV show strong bleach bands at their
ground-state frequencies and strong absorption bands to higher
energy. These bands shift to higher energy due to reduced
back-bonding from the Mo2δ orbital to the ligand π* orbitals in
S1 states. Also, the 3δδ* T1 state ν(CC) absorption bands of
II and IV are slightly asymmetric, probably due to ground-state
bleach.
fs-TRIR Spectra in the ν(C6)ring and ν(O−C−O)/ν(N−C−
O)/ν(N−C−S) Region. fs-TRIR spectra of I−IV from 1650 to
1350 cm−1 are shown in Figure 11. Strong features were
17
1
3
observed corresponding to the ν(C6)ring bands in the MLCT
their δδ* T1 states.
H
DOI: 10.1021/acs.inorgchem.6b02517
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Due to their closeness in energy, the νas(O−C−O) and
νas(N−C−O) features overlap in the spectra of I. The νas(O−
C−O)/νas(N−C−O) bleach band was observed at 1490 cm−1,
and the absorption band was observed at 1523 cm−1,
representing a 33 cm−1 shift to higher energy. The absorption
bands at 1415 and 1382 cm−1 were assigned to νs(O−C−O)
and νs(N−C−O) modes, respectively. In II, similar features
were observed to I. A bleach band at 1478 cm−1 and an
absorption band at 1516 cm−1 were attributed to νas(N−C−O),
while an absorption band at 1375 cm−1 was attributed to νs(N−
C−O). In III, a clear bleach band at 1484 cm−1 and an
absorption band at 1498 cm−1 were assigned to νas(O−C−O).
A set of weaker features at lower energy, including the bleach
band at 1403 cm−1 and the absorption band at 1416 cm−1, were
assigned to νas(N−C−S). In IV, a broad bleach band at 1423
cm−1 and an absorption feature at ∼1460 cm−1 were assigned
to νas(N−C−S).
compounds in both the ground and the MLCT excited states
using group theory. The ground-state CC stretch modes are
displayed in Figure 13. To simplify the derivation process, C2
Kinetic fitting of the fs-TRIR spectra is shown in Figure 12.
Kinetic traces were taken from the ν(CC) absorption bands
of I−IV, and the lifetimes of the 1MLCT S1 states of I−IV were
extracted.
Figure 13. CC stretching modes in the ground state of trans-
Mo2L2(O2C-TiPB)2 and Mo2L4 compounds. C2 and C4 point groups
were used for simplicity.
Traces of I and II were best fitted using biexponetial decay
functions. The first component (<1 ps) was attributed to
intramolecular cooling processes.16 III and IV were best fitted
using monoexponential decay functions. The S1 lifetimes of I−
IV were found to be 5.5, 6.9, 3.5, and 1.6 ps, respectively, in
close agreement with the fs-TA results.
and C4 point groups were used for the ground-state molecules.
These results align well with frequency analyses for I−IV in the
DFT calculations.
As shown in Figure 13, in the ground state of the trans-
Mo2L2(O2C-TiPB)2 compounds, the combination of CC
stretches gives two ν(CC) normal modes: the symmetric a
mode and the asymmetric b mode. The a mode is Raman
active, and the b mode is IR active. With an inversion center,
the Raman and IR modes here are exclusive. In the ground state
of the Mo2L4 compounds there are four ν(CC) modes,
namely, the a, b, and e modes where the symmetric a and b
modes are Raman active and the degenerate e modes are IR
active.
DISCUSSION
■
The ν(CC) shifts of compounds I−IV from fs-TRIR
experiments and from DFT calculations are summarized in
Table 5. As mentioned previously, we calculated the ν(CC)
Table 5. Shift of ν(CC) Bands in the 1MLCT S1 and 3δδ*
a
T1 States of I−IV
The MLCT-state CC stretching modes (with different
Δν(CC) S1 (cm−1
)
Δν(CC) T1 (cm−1
)
charge distribution patterns) are displayed in Figure 14. In the
1
3
experimental
computational
experimental
computational
MLCT excited states (including MLCT and MLCT states),
coupling of CC stretches depends on molecular geometry
and the molecular geometry in turn depends on charge
distribution.37,38 Thus, molecular distortions have to be
considered before the derivation of coupled CC stretch
modes. The ν(CC) normal modes in Figure 14 addressed all
extreme localized and delocalized cases.
I
−53, −235
−53, −235
−189
−153
−76, −104
−126
3
3
II
5
5
III
IV
−17
−15
−2
2
−160, −234
−74, −122
a
Notes: Computational details can be found in the Experimental
Section.
Figure 14A and 14B demonstrates the ν(CC) modes in
the delocalized and the localized MLCT states of the trans-
Mo2L2(O2C-TiPB)2 compounds. In the delocalized MLCT
state, the symmetry of the ground state is conserved and one
Raman active mode and one IR active mode are expected. In
the localized MLCT state, however, the molecule loses its
symmetric structure and there would be two IR active ν(CC)
modes. One of the IR active ν(CC) modes only involves the
neutral ligand and would reside close to the ground-state
frequency. The other mode would show a much greater shift
due to the extra electron density in the singly reduced ligand.
Figure 14C−F demonstrates the ν(CC) modes in the
MLCT states of Mo2L4 compounds. Overall, there are four
extreme scenarios of how the transferred electron can be
distributed over the four ligands. The distortions caused by the
electron density distribution can be regarded as results of the
excited-state Jahn−Teller effect mentioned earlier. Which exact
distortion scenario is invoked depends on the electronic
structure details.37
shifts in the 3δδ* T1 states (3I′−3IV′) and estimated the ν(C
1
C) shifts in the MLCT S1 states of I−IV using molecular
anions (I′−−IV′−) as mimics. The vibrational shifts predicted
by the molecular anion calculations correspond to completely
delocalized excited states where the transferred electron resides
on the two trans ligands in I and III and on all four ligands in II
and IV. Compared with experimental observations, the
3
computational results for the δδ* T1 states are reasonably
close, but the results from anion calculations for the 1MLCT S1
states are less accurate due to deviations from completely
delocalized states.
In the process of examining the computational results and
the fs-TRIR spectra, it becomes clear that both the coupling of
CC stretches and the excited-state charge distribution
account for the observed ν(CC) bands in Mo2 paddlewheel
compounds.
To better assign the observed ν(CC) bands, we derived all
CC normal modes in trans-Mo2L2(O2C-TiPB)2 and Mo2L4
I
DOI: 10.1021/acs.inorgchem.6b02517
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 14. CC stretching modes in the MLCT states of trans-Mo2L2(O2C-TiPB)2 compounds where (A) electron is delocalized and (B) electron
is localized and in the MLCT states of Mo2L4 compounds where (C) electron is delocalized over four ligands, (D) electron is delocalized over trans
ligands, (E) electron is delocalized over cis ligands, and (F) electron is localized on one ligand.
When the transferred electron is equally distributed over all
four ligands, the Jahn−Teller effect would lift the degeneracy of
the e orbitals which leads the dihedral angle between the cis
ligands in Mo2L4 compounds to contract or expand. This is
observed in the DFT calculation for the Mo2L4 anions (the
optimized structures of II′− and IV′− are shown in Figure S9).
The coupling of the CC stretches generates two e-like IR
active modes, each slightly different from the e modes of the
ground state as shown in Figure 14C. These two e-like modes
have produced the dual ν(CC) bands in the anion
calculations of II and IV (Table 5, Figures S11 and S13).
When the transferred electron is delocalized over only one
set of trans ligands, coupling of CC stretches generates two
IR active ν(CC) modes. One ν(CC) mode only involves
the neutral ligands, and the other only involves the reduced
ligands as shown in Figure 14D. In this scenario, the shift of
ν(CC) bands observed in a Mo2L4 compound would be
similar to a trans-Mo2L2(O2C-TiPB)2 compound with the same
acceptor ligands. When the transferred electron is delocalized
over a set of cis ligands, the coupling of CC stretches would
be in a similar way (shown in Figure 14E).
In the extreme localized case, the transferred electron resides
only on one of the four ligands. As shown in Figure 14F, the
coupling of CC stretches generates four IR active ν(CC)
modes. Three of them involve only neutral ligands, giving
overlap IR active modes close to the ground state. The
remaining ν(CC) mode involves the singly reduced ligand
and is expected to show a great shift from the ground state. The
1MLCT S1 state of compound Mo2[(NPh)2CCCPh]4
studied previously was thought to be totally localized in this
manner.24
It is clear that the group theory analyses are able to give
qualitative predictions of ν(CC) bands of the paddlewheel
compounds in the totally delocalized and totally localized
MLCT states. However, the analyses provide little information
on the energy and intensity of these IR active bands. The
assignments of charge distribution patterns in the MLCT states
of trans-M2L2L2 and M2L4 compounds have to be based on the
combination of computational and experimental results and
comparison between related compounds.
In addition to these extreme cases, as we mentioned before,
there are borderline cases lying between totally delocalized and
localized MLCT states. To better evaluate charge distribution
in the MLCT excited states, a mixed-valence classification
scheme similar to the Robin−Day scheme (from class I to class
III) has been used to classify them.16,39 Following previously
published works, a class I MLCT state is a charge-localized
MLCT state; a class III MLCT state is a charge-delocalized
MLCT state, and a class II MLCT state is an intermediate case
where the barrier for electron transfer between ligands is
relatively small. Since we are using infrared spectroscopy as the
characterization method, any MLCT excited state where the
electron transfer is much slower than the IR time scale (∼10−12
s) will be perceived as class I.40,41
In the fs-TRIR spectra of I, the appearance of two ν(CC)
1
bands in the MLCT S1 state (shifted by 53 and 235 cm−1)
excludes the possibility of the transferred electron being totally
delocalized according to the group theory analyses. The two
ν(CC) bands in I are similar to the dual ν(CC) bands
J
DOI: 10.1021/acs.inorgchem.6b02517
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
observed in compounds like Mo2[(NPh)2CCCPh]2(O2C-
TiPB)2 (shifted by 70 and 235 cm−1) and Mo2[(NiPr)2CC
CPh]2(O2CMe)2 (shifted by 45 and 241 cm−1).16,24 Consistent
with previous assignments, the appearance of two ν(CC)
bands is attributed to two unequally reduced amide ligands in
the MLCT S1 state of I, and thus, the MLCT S1 state of I is
assigned as class II.
LUMO of III as shown by the DFT calculation could have
promoted the electron transfer between the thioamide ligands.
Besides the differences between the amide and the thioamide
ligands, the arrangements of the ligands also appear to impact
on the interligand electron transfer. The amide ligands in
compounds I and II adopt the same centrosymmetric
arrangement, and electron transfer between the amide ligands
1
1
1
Compound II displayed two ν(CC) bands similar to I in
in the MLCT S1 state of I and II also appears to be similar.
1
The supporting O2C-TiPB ligands showed limited influence on
the charge distribution in I. The charge distribution differences
in III and IV were then attributed to the distinctly
arrangements of the thioamide ligands. The centrosymmetric
arrangement of the thioamide ligands in III appears to favor
faster interligand electron transfer in comparison to the mirror-
symmetric arrangement in IV. Being of different point groups,
it is hard to directly compare the electronic structures of III and
IV. The LUMO of III does show more metal character in
comparison to the LUMO+1 of IV, and this could be taken as
evidence of different degrees of metal−ligand orbital mixing
induced by ligand arrangements. The details of interligand
electron transfer process require further investigation.
the MLCT S1 state, and the shift magnitudes of the ν(CC)
bands are similar as well. Thus, it is reasonable to propose that I
and II have the same charge distribution pattern and the
1MLCT S1 state of II is also assigned as class II. With four
MLCT active ligands, the transferred electron in II appears to
be limited to one set of trans ligands, following the coupling
scheme in Figure 14D. The significant breadth of the ν(CC)
bands in II could be attributed to solvent interactions or
exchange processes between the two sets of trans ligands.22
There is debate over whether the appearance of two ν(C
C) bands (with one slightly shifted to lower energy and one
greatly shifted to lower energy) fit better with the class I
scenario. The class II assignment in I is preferred based on the
1
lack of a second ν(CC) band for the MLCT S1 state of
compound Mo2[(NPh)2CCCPh]4 studied previously, which
was assigned as class I.24
CONCLUSION
Consistent with our study of Mo2 amidinate compounds,16,24
the strong absorption of Mo2 amide and thioamide compounds
■
In compound III, the single sharp ν(CC) band observed
in the 1MLCT S1 state suggests that III has a totally delocalized
1MLCT S1 state. The observed ν(CC) shift of 189 cm−1 is
larger than the 124 cm−1 predicted by the molecular anion
calculations. However, the relatively large ν(CC) shift could
be attributed to the result of greater electron density
1
in the UV−vis region generates MLCT S1 excited states. The
substitution of nitrogen atoms by oxygen or sulfur atoms,
nevertheless, caused significant changes in the electronic
structures and photophysical properties. On the basis of
femtosecond time-resolved spectra, especially the observed
ν(CC) bands in the fs-TRIR spectra, the transferred electron
was found to have various distribution patterns over the amide
and thioamide ligands in the 1MLCT S1 states of the
paddlewheel compounds.
Using a mixed-valence classification scheme, we assigned the
charge distribution in the 1MLCT S1 states of amide
compounds I and II both as class II. The majority of the
transferred electron density is located on one amide ligand, but
some of the transferred electron density also presents on the
other amide ligand (trans to the first). In the thioamide
compounds, however, the transferred electron was found to be
totally delocalized over two trans-thioamide ligands in the
1MLCT S1 state of III and unequally shared by two trans-
thioamide ligands in the 1MLCT S1 state of IV. The 1MLCT S1
state of III is assigned as class III, and the 1MLCT S1 state of IV
is assigned as class II.
1
concentration on the CC units in the MLCT S1 state of
III relative to compounds like I (n.b. the ν(C6)ring bands in III
and IV are shifted much less in comparison to those in I and
II). The possibility that III could have a class I 1MLCT S1 state
(if only one IR active mode appears in fs-TRIR spectra) was
dismissed with the observation of the ν(CC) band shifted to
1
even lower energy (234 cm−1) in IV. The MLCT S1 state of
III was then assigned as class III.
The presence of two close-lying ν(CC) bands in IV is
intriguing. The number of ν(CC) bands fits with the charge
delocalization over four ligands scenario (Figure 14C) or a
partial delocalization over two trans ligands scenario (similar to
Figure 14D). The observed large shift magnitudes of the ν(C
C) bands (160 and 234 cm−1) clearly exclude the former case
(see computational results in Table 5). The single ν(CC)
band observed in III (shifted by 189 cm−1) lies between the
two ν(CC) bands observed in IV (shifted by 160 and 234
cm−1). Thus, the 1MLCT S1 state of IV was assigned as class II
The observed different ν(CC) band patterns were
1
attributed to different rates of interligand electron transfer in
with reference to the class III assignment of the MLCT S1
1
the MLCT S1 state of these Mo2 paddlewheel compounds.
state of III. The arrangement of the ν(CC) bands in III and
IV raises the further suggestion of a possible coalescence
process.40
The stronger metal−ligand orbital mixing induced by sulfur
atoms in thioamide compounds appears to have promoted the
interligand electron transfer which yielded more delocalized
1MLCT S1 states. Ligand arrangements were also shown to
influence the metal−ligand orbital mixing and interligand
1
The observation of the class II MLCT S1 state in I and the
1
class III MLCT S1 state in III suggests that the interligand
electron transfer should be much faster in III than in I. To
generate different ν(CC) patterns in the fs-TRIR spectra, the
electron transfer has to take place on the IR time scale.
The difference in interligand electron transfer can be
attributed to the electronic structures of I and III, especially
the HOMO and LUMO. The LUMO, in particular, represents
an approximate description of the molecular orbital occupied
by the transferred electron in the 1MLCT S1 state. The
significant contribution from metal orbitals (Mo2δ*) in the
1
electron transfer in the MLCT S1 states.
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.inorg-
chem.6b02517.
K
DOI: 10.1021/acs.inorgchem.6b02517
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Crystal structure refinement information on I and III,
Article
(11) Kiesz, M. D.; Hoekstra, R. M.; Chen, Y. T.; Telo, J. P.; Nelsen,
S. F.; Zink, J. I. Coupled States in Dinitrofluorene: Relationships
between Ground State and Excited State Mixed Valence. J. Phys. Chem.
A 2014, 118, 11490−11498.
(12) Barlow, S.; Risko, C.; Odom, S. A.; Zheng, S.; Coropceanu, V.;
́
Beverina, L.; Bredas, J. L.; Marder, S. R. Tuning Delocalization in the
NBO analysis of I−IV, nanosecond TA spectra of I−IV,
NMR and MALDI-TOF mass spectra of I−IV (PDF)
Crystallographic information (CIF)
(CIF)
Radical Cations of 1,4-Bis[4-(diarylamino)styryl]benzenes, 2,5-Bis[4-
(diarylamino)styryl]thiophenes, and 2,5-Bis[4-(diarylamino)styryl]-
pyrroles through Substituent Effects. J. Am. Chem. Soc. 2012, 134,
10146−10155.
(13) Mani, T.; Grills, D. C.; Newton, M. D.; Miller, J. R. Electron
Localization of Anions Probed by Nitrile Vibrations. J. Am. Chem. Soc.
2015, 137, 10979−10991.
(14) Chisholm, M. H.; Brown-Xu, S. E.; Spilker, T. F. Photophysical
Studies of Metal to Ligand Charge Transfer Involving Quadruply
Bonded Complexes of Molybdenum and Tungsten. Acc. Chem. Res.
2015, 48, 877−885.
(15) Alberding, B. G.; Chisholm, M. H.; Gustafson, T. L.; Liu, Y.;
Reed, C. R.; Turro, C. Photophysical Studies of trans-Bis-
(phenylethynyldiisopropylamidinato)bis(acetato)dimetal Complexes
Involving MM Quadruple Bonds Where M = Mo or W. J. Phys.
Chem. A 2010, 114, 12675−12681.
AUTHOR INFORMATION
Corresponding Author
*E-mail: jiang.627@osu.edu.
ORCID
Changcheng Jiang: 0000-0002-7674-6883
Present Addresses
■
†School of Chemistry, University of Nottingham, University
Park, Nottingham NG7 2RD, United Kingdom.
‡Department of Chemistry, Northwestern University, Evan-
ston, Illinois 60208, United States.
Notes
The authors declare no competing financial interest.
(16) Alberding, B. G.; Chisholm, M. H.; Gallucci, J. C.; Ghosh, Y.;
Gustafson, T. L. Electron Delocalization in the S1 and T1 Metal-to-
Ligand Charge Transfer States of trans-Substituted Metal Quadruply
Bonded Complexes. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 8152−
8156.
(17) Alberding, B. G.; Chisholm, M. H.; Gustafson, T. L. Detection
of the Singlet and Triplet MM δδ* States in Quadruply Bonded
Dimetal Tetracarboxylates (M = Mo, W) by Time-Resolved Infrared
Spectroscopy. Inorg. Chem. 2012, 51, 491−498.
(18) Alberding, B. G.; Chisholm, M. H.; Chou, Y.-H.; Gallucci, J. C.;
Ghosh, Y.; Gustafson, T. L.; Patmore, N. J.; Reed, C. R.; Turro, C.
Quadruply Bonded Dimetal Units Supported by 2,4,6-Triisopropyl-
benzoates MM(TiPB)4 (MM = Mo2, MoW, and W2): Preparation and
Photophysical Properties. Inorg. Chem. 2009, 48, 4394−4399.
(19) Alberding, B. G.; Brown-Xu, S. E.; Chisholm, M. H.; Gustafson,
T. L.; Reed, C. R.; Naseri, V. Photophysical Properties of MM
Quadruply Bonded Complexes (M = Mo, W) Supported by
Carboxylate Ligands: Charge Delocalization and Dynamics in S1 and
T1 States. Dalton Trans. 2012, 41, 13097−13104.
(20) Brown-Xu, S. E.; Chisholm, M. H.; Durr, C. B.; Spilker, T. F.
Concerning the Ground State and S1 and T1 Photoexcited States of
the Homoleptic Quadruply Bonded Complexes Mo2(O2CC6H4-p-X)
4, Where X = CC−H or CN. J. Phys. Chem. A 2013, 117, 13893−
13898.
(21) Chisholm, M. H.; Lear, B. J. M2δ to Ligand π-Conjugation:
Testbeds for Current Theories of Mixed Valence in Ground and
Photoexcited States of Molecular Systems. Chem. Soc. Rev. 2011, 40,
5254−5265.
(22) Stark, C. W.; Schreier, W. J.; Lucon, J.; Edwards, E.; Douglas, T.;
Kohler, B. Interligand Electron Transfer in Heteroleptic Ruthenium-
(II) Complexes Occurs on Multiple Time Scales. J. Phys. Chem. A
2015, 119, 4813−4824.
(23) Bunting, P.; Chisholm, M. H.; Gallucci, J. C.; Lear, B. J. Extent
of M2δ to Ligand π-Conjugation in Neutral and Mixed Valence States
of Bis(4-isonicotinate)-bis(2,4,6-triisopropylbenzoate) Dimetal Com-
plexes (MM), Where M = Mo or W, and Their Adducts with
Tris(pentafluorophenyl)boron. J. Am. Chem. Soc. 2011, 133, 5873−
5881.
ACKNOWLEDGMENTS
■
We would like to thank the National Science Foundation for
the funding associated with grant numbers CHE-0957191 and
CHE-1266298. We are grateful to the Ohio State University
Center for Chemical and Biophysical Dynamics for use of the
laser systems, the Ohio Supercomputer Center for computa-
tional resources, Prof. Claudia Turro and Prof. Ewan Hamilton
for useful discussion, and Prof. Prabir Dutta for use of
instrumentation.
REFERENCES
■
344.
(1) Gratzel, M. Photoelectrochemical Cells. Nature 2001, 414, 338−
̈
(2) Chan, S. C.; Chan, M. C. W.; Wang, Y.; Che, C. M.; Cheung, K.
K.; Zhu, N. Organic Light-Emitting Materials Based on Bis-
(arylacetylide)platinum(II) Complexes Bearing Substituted Bipyridine
and Phenanthroline Ligands: Photo- and Electroluminescence from
3MLCT Excited States. Chem. - Eur. J. 2001, 7, 4180−4190.
(3) Morris, A. J.; Meyer, G. J.; Fujita, E. Molecular Approaches to the
Photocatalytic Reduction of Carbon Dioxide for Solar Fuels. Acc.
Chem. Res. 2009, 42, 1983−1994.
(4) Sattler, W.; Ener, M. E.; Blakemore, J. D.; Rachford, A. A.;
LaBeaume, P. J.; Thackeray, J. W.; Cameron, J. F.; Winkler, J. R.; Gray,
H. B. Generation of Powerful Tungsten Reductants by Visible Light
Excitation. J. Am. Chem. Soc. 2013, 135, 10614−10617.
(5) Demadis, K. D.; Hartshorn, C. M.; Meyer, T. J. The Localized-to-
Delocalized Transition in Mixed-Valence Chemistry. Chem. Rev. 2001,
101, 2655−2686.
(6) Yeh, A. T.; Shank, C. V.; McCusker, J. K. Ultrafast Electron
Localization Dynamics Following Photo-Induced Charge Transfer.
Science 2000, 289, 935−938.
(7) Zhang, W.; Ji, M.; Sun, Z.; Gaffney, K. J. Dynamics of Solvent-
Mediated Electron Localization in Electronically Excited
Hexacyanoferrate(III). J. Am. Chem. Soc. 2012, 134, 2581−2588.
(8) Hupp, J. T.; Williams, R. D. Using Resonance Raman
Spectroscopy To Examine Vibrational Barriers to Electron Transfer
and Electronic Delocalization. Acc. Chem. Res. 2001, 34, 808−817.
(24) Jiang, C.; Young, P. J.; Durr, C. B.; Spilker, T. F.; Chisholm, M.
H. Synthesis, Structure, and Photophysical Properties of Mo2(NN)4
and Mo2(NN)2(TiPB)2, Where NN = N,N′-Diphenylphenylpropiola-
midinate and TiPB = 2,4,6-Triisopropylbenzoate. Inorg. Chem. 2016,
55, 5836−5844.
(25) In Multiple Bonds between Metal Atoms, 3rd ed.; Cotton, F. A.,
Murillo, C. A., Walton, R. A., Eds.; Springer Science and Business
Media: New York, 2005.
(9) Kvapilova,
P.; Towrie, M.; Gray, H. B.; Zal
́
H.; Sattler, W.; Sattler, A.; Sazanovich, I. V.; Clark, I.
is, S.; Vlcek, A. Electronic Excited
́
̌
̌
States of Tungsten(0) Arylisocyanides. Inorg. Chem. 2015, 54, 8518−
8528.
(10) Dereka, B.; Rosspeintner, A.; Li, Z.; Liska, R.; Vauthey, E. Direct
Visualization of Excited-State Symmetry Breaking Using Ultrafast
Time-Resolved Infrared Spectroscopy. J. Am. Chem. Soc. 2016, 138,
4643−4649.
L
DOI: 10.1021/acs.inorgchem.6b02517
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(26) Hicks, J.; Ring, S. P.; Patmore, N. J. Tuning the Electronic
Structure of Mo−Mo Quadruple Bonds by N for O for S Substitution.
Dalton Trans. 2012, 41, 6641−6650.
(27) Armarego, W. L. F.; Chai, C. L. L. In Purification of Laboratory
Chemicals, 5th ed.; Butterworth-Heinemann: Amsterdam, Boston,
2003.
(28) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.
P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;
Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,
K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.;
Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega,
N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;
Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.;
Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.;
̈
Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. In Gaussian
09; Gaussian, Inc.: Wallingford, CT, 2009.
(29) Andrae, D.; Haußermann, U.; Dolg, M.; Stoll, H.; Preuß, H.
̈
Energy-Adjusted ab initio Pseudopotentials for the Second and Third
Row Transition Elements. Theor. Chim. Acta 1990, 77, 123−141.
(30) Irikura, K. K.; Johnson, R. D.; Kacker, R. N. Uncertainties in
Scaling Factors for ab initio Vibrational Frequencies. J. Phys. Chem. A
2005, 109, 8430−8437.
(31) Burdzinski, G.; Hackett, J. C.; Wang, J.; Gustafson, T. L.; Hadad,
C. M.; Platz, M. S. Early Events in the Photochemistry of Aryl Azides
from Femtosecond UV/Vis Spectroscopy and Quantum Chemical
Calculations. J. Am. Chem. Soc. 2006, 128, 13402−13411.
(32) Wang, J.; Burdzinski, G.; Kubicki, J.; Platz, M. S. Ultrafast UV−
Vis and IR Studies of p-Biphenylyl Acetyl and Carbomethoxy
Carbenes. J. Am. Chem. Soc. 2008, 130, 11195−11209.
(33) Dolinar, B. S.; Berry, J. F. Electronic Tuning of
Mo2(thioamidate)4 Complexes through π-System Substituents and
cis/trans Isomerism. Dalton Trans 2014, 43, 6165−6176.
(34) Glendening, E. D.; Reed, A. E.; Carpenter, J. E.; Weinhold, F.
NBO, Version 3.1.
(35) Lu, T.; Chen, F. Multiwfn: A Multifunctional Wavefunction
Analyzer. J. Comput. Chem. 2012, 33, 580−592.
(36) Norman, J. G.; Kolari, H. J.; Gray, H. B.; Trogler, W. C.
Electronic Structure of Dimolybdenum Tetraformate, Dimolybdenum-
(4+) Ion, and Dimolybdenum. Inorg. Chem. 1977, 16, 987−993.
(37) Bersuker, I. B. The Jahn-Teller Effect and Vibronic Interactions in
Modern Chemistry; Modern Inorganic Chemistry; Plenum Press: New
York, 1984.
(38) Scattergood, P. A.; Delor, M.; Sazanovich, I. V.; Bouganov, O.
V.; Tikhomirov, S. A.; Stasheuski, A. S.; Parker, A. W.; Greetham, G.
M.; Towrie, M.; Davies, E. S.; et al. Electron Transfer Dynamics and
Excited State Branching in a Charge-Transfer Platinum(II) Donor−
bridge-Acceptor Assembly. Dalton Trans 2014, 43, 17677−17693.
(39) Plummer, E. A.; Zink, J. I. Excited-State Mixed Valence in
Transition Metal Complexes. Inorg. Chem. 2006, 45, 6556−6558.
(40) Ito, T.; Hamaguchi, T.; Nagino, H.; Yamaguchi, T.; Kido, H.;
Zavarine, I. S.; Richmond, T.; Washington, J.; Kubiak, C. P. Electron
Transfer on the Infrared Vibrational Time Scale in the Mixed Valence
State of 1,4-Pyrazine- and 4,4′-Bipyridine-Bridged Ruthenium Cluster
Complexes. J. Am. Chem. Soc. 1999, 121, 4625−4632.
(41) Rocha, R. C.; Shreve, A. P. Exploring the Localized-to-
Delocalized Boundary in Mixed-Valence Systems Using Infrared
Spectroelectrochemistry. Inorg. Chem. 2004, 43, 2231−2233.
M
DOI: 10.1021/acs.inorgchem.6b02517
Inorg. Chem. XXXX, XXX, XXX−XXX