Near-infrared to violet triplet-triplet annihilation fluorescence upconversion of Os(ii) complexes by strong spin-forbidden transition

Article abstract of DOI:10.1039/c9dt02276g

Three Os(ii) complexes were synthesized with ligands 2,2′-dipyridyl (dipy), 1,10-phenanthroline monohydrate (phen), and 4,7-diphenyl-1,10-phenanthroline (diphen), and applied as triplet photosensitizers for triplet-triplet annihilation (TTA) fluorescence

Full text of DOI:10.1039/c9dt02276g

Dalton
Transactions
View Article Online
View Journal
PAPER
Near-infrared to violet triplet–triplet annihilation
fluorescence upconversion of Os(II) complexes by
strong spin-forbidden transition†
Cite this: DOI: 10.1039/c9dt02276g
a
Yaxiong Wei,a Min Zheng,a Lin Chen,*b Xiaoguo Zhou *a and Shilin Liu
Three Os(II) complexes were synthesized with ligands 2,2’-dipyridyl (dipy), 1,10-phenanthroline mono-
hydrate (phen), and 4,7-diphenyl-1,10-phenanthroline (diphen), and applied as triplet photosensitizers for
triplet–triplet annihilation (TTA) fluorescence upconversion. The strong spin-orbital coupling made direct
spin-forbidden transition of S0–T1 feasible. Lifetimes of the lowest triplet state of these complexes were
determined to be 107 ns, 373 ns, and 386 ns for Os-dipy, Os-phen, and Os-diphen, respectively, using
nanosecond transient absorption spectra. From steady-state phosphorescence emission spectra, energies
of the triplet states were derived to be 1.75 eV, 1.80 eV, and 1.74 eV for Os-dipy, Os-phen, and Os-diphen,
respectively. Using these photosensitizers, strong upconverted fluorescence of the triplet acceptors,
9,10-diphenylanthracene (DPA), perylene, and 9,10-bis(phenethynyl) anthracene (BPEA), was observed in
the visible to violet range. In particular, fluorescence emission with the largest anti-Stokes shift of 1.14 eV
was observed for the Os-phen/DPA system, and the upconverted quantum yield was determined as 5.9%
in deoxygenated dichloroethane. Additionally, upconversion was determined in air using mixtures of
dichloroethane and DMSO solvents, and the maximal quantum yield was measured to be 4.5% for Os-
phen/DPA.
Received 29th May 2019,
Accepted 7th July 2019
DOI: 10.1039/c9dt02276g
rsc.li/dalton
For light harvesting in a TTA upconversion system, a large
number of triplet photosensitizers have been synthesized.12–16
Introduction
Photon upconversion is a process in which the material pro- The photosensitizers based on transition metals or heavy
duces higher energy photons on irradiation with lower energy atoms such as Ir(II), Pt(II), Ru(II), I, and Br, are particularly
photons.1 Among the various well-known upconversion popular because of efficient intersystem crossing (ISC)
approaches, triplet–triplet annihilation (TTA) upconversion induced by strong spin–orbit coupling.3,7,17,18 Several organic
fluorescence is especially promising because of its low power compounds, e.g., 2,3-butanedione and benzophenone, also
density requirement (only a few mW cm−2, less than the solar exhibit efficient ISC because of a small energy gap between the
power density of ∼100 mW cm−2 at AM 1.5) and a relatively S1 and T1 states in low-lying n–π* transitions.19 Yet another
high efficiency under weak irradiation intensity.2–4 In the past type of photosensitizers are the fullerene family, i.e., C60 and
decades, TTA upconversion has been applied in numerous C70, since they are common intramolecular spin converters
fields, including photovoltaics,5,6 photoelectrochemistry,7 with ISC efficiency of unity.20–24 Recently, radical-enhanced
photocatalysis,8,9 and fluorescent cell imaging.10,11
ISC and spin–orbit charge transfer have been used to design a
novel triplet photosensitizer.25 However, taking into account
the energy loss from S1–T1, the anti-Stokes shift of upconverted
fluorescence has usually been reduced to less than 0.8 eV.13
The anti-Stokes shift is one of the most important para-
meters of a TTA upconversion system for solar energy appli-
cation. To utilize solar light efficiently, a large anti-Stokes shift
of upconverted fluorescence is necessary. A creative strategy to
increase the anti-Stokes shift involves minimizing the energy
loss of ISC in the formation of triplet photosensitizers, such as
thermally activated delayed fluorescence (TADF),26,27 or direct
singlet-to-triplet (S–T) excitation.28–30 Generally, TADF mole-
cules have a small S1–T1 gap, but the related absorption is
aHefei National Laboratory for Physical Sciences at the Microscale, iChEM
(Collaborative Innovation Center of Chemistry for Energy Materials), Department of
Chemical Physics, University of Science and Technology of China, Hefei,
Anhui 230026, China. E-mail: xzhou@ustc.edu.cn
bSchool of Physics and Materials Engineering, Hefei Normal University, Hefei,
Anhui 230601, China. E-mail: chenlin@hfnu.edu.cn
†Electronic supplementary information (ESI) available: Experimental apparatus
and measurement procedures; NMR spectra; transient absorption spectra and
dynamic decay curves; TTA upconversion spectra; absorption and fluorescence
spectra of Os(II) complexes; detailed DFT calculation results. See DOI: 10.1039/
c9dt02276g
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
always located in the blue and violet spectral regions.31,32
Compared to the TADF method, direct S–T excitation is more
practical because the absorptions are located in the near-
infrared region.33,34 However, the direct S–T excitation is spin-
forbidden for most compounds, so that the molar absorption
coefficients are extremely small (<10 M−1 cm−1).28 In 2005,
Altobello et al.33 reported a terpyridyl Os(II) complex as the first
successful photosensitizer of direct S–T excitation. The S–T
absorption at 1100 nm was accompanied with metal-to-ligand
charge transfer (MLCT). Recently, Amemori et al.30 applied
other Os(II) complexes as triplet photosensitizers in the TTA
upconversion of Rubrene, and the anti-Stokes shift was pro-
moted up to 0.86 eV with quantum yield (ΦUC) of 3.1% in solid
films. However, the triplet–triplet energy transfer (TTET)
efficiency and the quantum yield were found to be significantly
low in solutions as the triplet lifetime was only 12 ns. The
same authors have also reported a new triplet photosensitizer,
Os(bptpy)22+, for direct S–T absorption later.29 According to
the extended triplet lifetime (207 ns), the anti-Stokes shift was
further increased to 0.97 eV, and the ΦUC was 2.7% in N,N-di-
methylformamide (DMF) solutions. Very recently, Liu et al.28
synthesized an Os(II)-tris(bpy) complex by attaching a Bodipy
moiety. The triplet state lifetime was markedly extended to
1.73 μs. The ΦUC value was measured to be 1.2% with N,N′-bis
(3-pentyl)perylene-3,4,9,10-bis(dicarboximide) (PBI) as an
energy acceptor, however, the anti-Stokes shift was only 0.3 eV.
In the TTA upconversion mechanism, the TTET process is
usually rate-determining.35 Thus, a brief lifetime of the triplet
photosensitizer prepared by direct S–T transition significantly
diminishes TTET efficiency29,30 and leads to unsatisfying per-
formances. Moreover, the relatively large energy gaps between
the triplet states of photosensitizers and acceptors cause extra
energy loss in the TTET process.28,29
Using different ligands, the lowest triplet energies of the
complexes can be altered purposefully.36 Thus, we synthesized
three Os(II) complexes, namely, Os-dipy, Os-phen, and Os-
diphen, with 2,2′-dipyridyl (dipy), 1,10-phenanthroline mono-
hydrate (phen), and 4,7-diphenyl-1,10-phenanthroline (diphen),
respectively, to prolong the lifetimes of triplet photosensitizers
and find consistent partners of TTET to reduce the extra energy
loss. Among the three photosensitizers, Os-dipy was reported,
however its performance in TTA upconversion application was
unsatisfied (the anti-Stokes shift was 0.3 eV and ΦUC was
1.2%).28 Scheme 1 shows the molecular structures of these
triplet photosensitizers and acceptors. Using steady-state and
transient absorption/emission spectroscopy, the photophysical
properties of these Os(II) complexes, such as lifetimes and exci-
tation energies of the triplet states, were investigated. Based on
the triplet energies of the complexes, DPA, perylene, and BPEA
were chosen as triplet acceptors. According to the direct S–T
excitation and the matched triplet energies, a larger anti-Stokes
shift (1.14 eV) than the reported values was achieved in the TTA
upconversion system involving Os(II)-phen and DPA in solu-
tions. Moreover, the mixed solvent of DMSO and dichloroethane
was tested for these upconversion systems, to find a simple and
feasible way to apply them in air.
Scheme 1 Molecular structures of the photosensitizers, Os-dipy, Os-
phen, Os-diphen, and the energy acceptors, 9,10-Diphenylanthracene
(DPA), perylene, and 9,10-bis(2-phenylethynyl)-anthracene (BPEA).
Results and discussion
Steady-state UV-Vis absorption and phosphorescence spectra
The UV-Vis absorption spectra of the photosensitizers in
dichloroethane are shown in Fig. 1a. The Os(II) complexes
exhibited similar spectral absorptions, involving a doublet
absorption in the region of 400–550 nm and a weak peak at
∼660 nm with a long-wavelength tail. For Os-phen in dichlor-
oethane, two intense singlet–singlet absorption peaks were
observed at 432 nm (ε = 29 000 M−1 cm−1) and 480 nm (ε =
26 000 M−1 cm−1), both of which were assigned as the singlet–
singlet MLCT transitions. In contrast, the weak and broad
absorption of singlet–triplet MLCT transition was also identi-
fied, although dim double peaks were observed. For Os-
diphen, the double singlet–singlet absorptions were located at
450 nm (ε = 41 000 M−1 cm−1) and 501 nm (ε = 35 000
M−1 cm−1), while the singlet–triplet transition peak covered a
range of 560–720 nm with a molar absorption coefficient of
7300 M−1 cm−1 at 682 nm. For Os-dipy, more absorption peaks
were observed in the wavelength region. The singlet–singlet
absorptions were observed at 370 nm (ε = 14 000 M−1 cm−1),
436 nm (ε = 17 000 M−1 cm−1), and 483 nm (ε = 17 000
M−1 cm−1), while the singlet–triplet band was located at
Fig. 1 (a) Steady-state UV-Vis absorption spectra of Os-dipy, Os-phen,
and Os-diphen. (b) Phosphorescence emission spectra (λex = 663 nm) of
Os-dipy, Os-phen, and Os-diphen. Dichloroethane as solvent, c = 1 ×
10−5 M, 25 °C.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
Table 2 bimolecular quenching rate constants (in 109 M−1 s−1) of
different acceptorsa
667 nm (ε = 4400 M−1 cm−1). Moreover, the concentration-
dependencies of absorption spectra of the Os(II) complexes
were studied (see ESI, Fig. S1–S4†). The absorption
intensity changed linearly with the concentration of the com-
plexes, indicating that no aggregation of the complexes
occurred in solution within the experimental concentration
range.
Compounds
Os-dipy
Os-phen
Os-diphen
–
DPA
τ0
τ
ΦTTET
kq
107
70
0.35
0.39
373
67
0.82
0.82
386
208
0.46
0.16
The fluorescence emission spectra of the complexes were
measured (see ESI, Fig. S5†) under photoexcitation at 450 nm
(singlet–singlet absorption). The fluorescence emission from
the singlet electronically excited state was very weak in the
region of 500–550 nm, implying a high ISC efficiency. In con-
trast to fluorescence, the phosphorescence emissions of these
complexes were very evident. Fig. 1b shows the phosphor-
escence spectra of the complexes in dichloroethane under
photoexcitation at 663 nm (direct S–T transition). For Os-dipy,
the emission covered a broad region of 600–900 nm with a
peak at 707 nm, and thus the excitation energy of the triplet
perylene
BPEA
τ
45
0.58
3.01
56
0.85
2.99
93
0.76
1.89
ΦTTET
kq
τ
69
0.36
1.89
112
0.7
1.79
198
0.49
0.96
ΦTTET
kq
a c[Photosensitizer] = 1 × 10−5 M, c[DPA] = 1.5 × 10−2 M, c[Pery] = 5 ×
10−3 M, c[BPEA] = 3.3 × 10−3 M.
state was determined as 1.75 eV. Owing to quenching by efficiency, their following kinetic behaviours had no difference
oxygen molecules, the phosphorescence from the triplet state after photoexcitation along singlet–singlet or singlet–triplet
was weak in air, but more than doubled under argon. The transitions. Thus, for convenience, the nanosecond time-
phosphorescence quantum yield was determined as 0.4% in resolved measurements were performed at excitation at
air and 0.9% in argon. For Os-phen in dichloroethane, the 532 nm. The lifetimes of the triplet state, τ0, TTET efficiencies
emission peak was located at 690 nm, and hence the excitation from photosensitizers to acceptors, ΦTTET, and the bimolecular
energy of the triplet state was 1.80 eV. Compared to Os-dipy, quenching rate constants, kq, were obtained and listed in
Os-phen showed stronger phosphorescence emission, and the Table 2.
quantum yield was increased to 2.0% in air and 5.5% in
Since the three complexes showed similar spectra and
argon. When two phenyls were attached to 1,10-phenanthro- kinetic behaviors, only the transient absorption spectra and
line, the extent of conjugation in the Os(II) complex was kinetic curves of Os-phen have been shown in Fig. 2a and b as
further enhanced in Os-diphen. As a result, the excitation an example. For Os-phen in dichloroethane, a very strong
energy of the triplet state was lowered to 1.74 eV, since the phosphorescence emission at 728 nm was observed together
phosphorescence emission peak was red-shifted to 711 nm. with two weak ground-state bleaching (GSB) peaks at 444 nm
However, the quantum yield was almost unchanged. Table 1 and 494 nm. Owing to its extremely weak intensity, the triplet
summarizes the photophysical properties of the three Os(II) excited state absorption (ESA) could not be clearly observed in
complexes in dichloroethane.
the range of 400–850 nm. Thus, the lifetime of the triplet state
of Os-phen was derived from the decay rate of phosphor-
escence emission as 373 ns (Fig. 2b). For Os-dipy, the phos-
phorescence emission of the triplet state was observed at
725 nm, accompanied with a weak GSB at 450 nm (see ESI,
Fig. S6†). The lifetime of the triplet Os-dipy was as short as
107 ns (see ESI, Fig. S7†), which was consistent with the
previous data.28 For Os-diphen, the photophysical features of
the triplet state were very similar to those of Os-phen. Two
GSB bands were observed at 460 nm and 510 nm, in addition
to a phosphorescence emission at 727 nm, and the triplet
state lifetime was determined as 386 ns (see ESI, Fig. S8
and S9†).
Transient absorption spectra and quenching rate constant
In the presence of triplet acceptors, TTET can occur once the
triplet state of the photosensitizer is prepared. Nanosecond
transient absorption spectra and kinetic curves of the charac-
teristic absorptions were performed for Os-dipy, Os-phen, and
Os-diphen. Since ISC of these photosensitizers were very
Table 1 Photophysical properties of the Os(II) complexes in
dichloroethanea
ΦP
ΦP
λabs ε (t)c λem (air)d (Ar)d
In the presence of triplet acceptors, TTET from the triplet
photosensitizers is feasible. Considering the T1 energies of the
photosensitizers in dichloroethane, DPA (T1 = 1.77 eV), peryl-
ene (T1 = 1.53 eV), and BPEA (T1 = 1.25 eV) were chosen as the
triplet acceptors in this study. As shown in Fig. 2b, the triplet
lifetime of Os-phen was reduced in the presence of the accep-
tors. Through fitting the decay curve of the absorption at
720 nm at different concentrations of the acceptors, the appar-
ent lifetime of the triplet complex, τ, was obtained. Using the
b
b
b
Compound λabs
ε (s)c
Os-dipy
Os-phen
370/436/483 1.4/1.7/1.7 667 0.44 707 0.4
0.9
5.5
4.4
432/480
Os-diphen 450/501
2.9/2.6
4.1/3.5
654 0.53 690 2.0
682 0.73 711 2.2
a c[Complex] = 1.0 × 10−5 M. b Peak positions of absorption and emis-
sion spectra, in unit of nm. c Molar extinction coefficient at maximal
absorption, ‘s’ is singlet excitation, ‘t’ is triplet excitation, in 104 M−1
cm−1 d Phosphorescence quantum yield, using methylene blue as a
.
standard (ΦFL = 3%).
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Fig. 2 (a) Nanosecond transient absorption spectra of Os-phen. (b) Decay curve of Os-phen at 720 nm in the presence of different acceptors. (c)
Energy diagram of the lowest triplet states of photosensitizers and acceptors. (d–f) Bimolecular quenching rates of the triplet photosensitizers by
different concentrations of the acceptors. c[Photosensitizer] = 1 × 10−5 M, in dichloroethane, λex = 532 nm, 25 °C.
Stern–Volmer eqn (1), bimolecular quenching rate constant, of the complexes were much lower than kdiffuse. Thus, TTET in
kq, was determined (Fig. 2d–f), and TTET efficiency, ΦTTET, was a solvent cage is the rate-determining step.
calculated using eqn (2).
As shown in Fig. 2c, the triplet energies of Os-phen and DPA
were most proximate, and therefore, they were the best energy
partners. The slightly higher energy of triplet Os-phen (1.80 eV)
than that of DPA (1.77 eV) facilitated the TTET process. The
triplet energies of Os-dipy and Os-diphen were slightly lower, i.e.,
1.75 eV and 1.74 eV, respectively. However, the mean thermal
activated energy at room temperature (25 °C) was estimated as
RT = 0.026 V, and thus, TTETs from these two Os(II) complexes
to DPA were still feasible, albeit with slower rates. When perylene
and BPEA were used as acceptors, all the Os(II) complexes had
enough active energy to perform TTET to form triplet acceptors,
but the energy loss was larger (e.g., 0.25 eV from Os-phen to pery-
lene, 0.55 eV from Os-phen to BPEA, 0.49 eV from Os-diphen to
BPEA). In summary, all the kq values agreed with the theoretical
deductions based on the TTET process.
1=τ ꢀ 1=τ0 ¼ kq ½acceptorꢁ
ð1Þ
ð2Þ
ΦTTET ¼ ðkTTET ꢂ ½acceptorꢁÞ=ðkTTET ꢂ ½acceptorꢁ þ kNR
¼ ð1=τ ꢀ 1=τ0Þ=ð1=τÞ
Þ
where τ0 and τ are the triplet lifetimes of the Os(II) complexes
in the absence and presence of acceptors, respectively. The kq
and ΦTTET values of the photosensitizers quenched by
different acceptors are summarized in Table 2. For DPA, kq
showed a remarkable dependence on the type of the photosen-
sitizer, e.g., 0.39 × 109 M−1 s−1 for Os-dipy, 0.82 × 109 M−1 s−1
for Os-phen, and 0.16 × 109 M−1 s−1 for Os-diphen, although
the molecular structures of the complexes were not dissimilar.
In the case of Os-dipy and Os-diphen, similar results were
observed with the three acceptors (see ESI, Fig. S7 and S9†).
DFT calculations on the frontier molecular orbitals
In principle, the apparent TTET rate between the photosen- To gain further insight into the photophysical features of the
sitizer and acceptor in solutions is determined by the for- Os(II) complexes, DFT calculations of their frontier molecular
mation rate of contact pair by diffusion and the TTET orbitals were performed.37 The excitation energies, oscillator
efficiency in a solvent cage. The former can be calculated as strengths, and main configurations of the low-lying electronic
the diffusion rate, while the latter generally depends on the excited states of the Os(II) complexes are summarized in Tables
energy gap between the triplet photosensitizer and acceptor. A S1–S3 of ESI.† For instance, in the case of Os-phen, the domi-
small energy loss from the triplet photosensitizer to the accep- nant configurations of the UV-Vis absorptions and the energy
tor implies higher TTET efficiency. In the present experiments, gap between the lowest triplet state and ground state can be
the diffusion rate, kdiffuse, was theoretically determined as seen in Fig. 3a. All the major absorptions in the UV-Vis range
8kBT/3η = 8.1 × 109 M−1 s−1 (where η is the viscosity of dichloro- were accompanied with MLCT. The other two complexes dis-
ethane, 0.84 mPa s at 25 °C). It is evident that all the kq values played similar characteristics (see ESI, Fig. S10 and S11†).
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
Fig. 4 (a–c) Upconverted fluorescence emission spectra under
different excitation power densities, where Os-phen was the photosen-
sitizer, and DPA (2.0 × 10–4 M), perylene (2.0 × 10–4 M), and BPEA (3.3
× 10–4 M) were the acceptors. (d–f) Double logarithmic plot of the
upconverted fluorescence intensity as a function of excitation power
density. Dichloroethane as the solvent, c[photosensitizer] = 1 × 10–5 M,
λex = 663 nm, 25 °C.
Fig. 3 (a) Frontier molecular orbitals involved in the low-lying singlet
and triplet excited states of the photosensitizers. (b) Spin density sur-
faces obtained by optimizing the triplet excited states. Calculations were
performed at the B3LYP/6-31G(d) level. Acetonitrile was used as the
solvent.
rescence spectra were obtained in deoxygenated dichlor-
oethane, as shown in Fig. S12–S14 of ESI.†
According to the DFT calculations, the energy differences of
HOMO–LUMO were determined to be 2.24 eV for Os-dipy, 2.30
eV for Os-phen, and 2.23 eV for Os-diphen, which shows the
same trend as the Redox potential results (see ESI Table S4†).
However, the UV-visible absorption spectra are greatly broad-
ened mainly owing to the thermal distribution of molecular
vibration, leading to that the HOMO–LUMO energy intervals
derived from the absorption spectra have relatively large uncer-
tainty (see ESI Table S4†). Fig. 3b shows the spin density sur-
faces of the lowest triplet states of the complexes. In the triplet
states, the spin densities were mainly distributed on the Os(II)
atom and one ligand. Thus, TTET was inevitably restricted by
steric hindrance in a solvent cage, leading to a lower kq value
than kdiffuse. Particularly, for Os-diphen, ligand inhibition was
more obvious, resulting in the lowest kq value among the three
complexes (Table 2).
In the present TTA upconversion system, four kinetic sub-
processes are involved, such as spin-forbidden absorption,
TTET to triplet acceptors, TTA, and fluorescence emission of
the acceptors. Thus, to achieve higher TTA upconversion
quantum yield, a high concentration of the acceptor is usually
recommended to saturate the TTET process. However, the self-
absorption of the acceptor at too high concentrations may
reduce the fluorescence quantum yield. In comparison with
the fluorescence spectrum of DPA itself (Fig. 5d), the band
intensity at 417 nm in the upconverted fluorescence spectrum
of Fig. 4a was reduced in some extent relative to the other
nearby peaks, confirming the self-absorption effect. Thus, it
was necessary to assess the effect of acceptor concentration on
Φ
UC in the experiments.
Fig. 5a–c show the dependence of the upconverted fluo-
rescence intensity on the acceptor concentration. For Os-dipy
and Os-diphen, the upconverted fluorescence intensity of DPA
was linearly enhanced with its concentration, implying that
the TTET stage was rate-determining in the overall TTA upcon-
version process. For Os-phen, the fluorescence intensity was
quickly increased and then gradually became saturated at ∼1.0
× 10−2 M. A similar phenomenon was observed for BPEA as the
acceptor, although the saturation concentration was slightly
reduced. However, perylene showed a different relationship
TTA upconversion fluorescence spectra in deoxygenated
dichloroethane
The upconverted fluorescence spectra of the triplet acceptors
when they were photoexcited at 663 nm in deoxygenated Os-
phen/dichloroethane solution are shown in Fig. 4a–c.† When
DPA was used as the acceptor, blue and violet emission was
clearly observed and exhibited an anti-Stokes shift of 1.14 eV.
For photosensitizers Os-dipy and Os-diphen, similar fluo-
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Table 3 ΦUC of the present TTA upconversion systemsa
ΦUC (dichloroethane,
Ar, %)b
ΦUC (DMSO, Air, %)c
Compound
DPA
Perylene
BPEA
DPA
Perylene
BPEA
d
Os-dipy
Os-phen
Os-diphen
0.4
5.9
1.2
1.1
2.3
2.5
3.0
12.5
10.3
—
0.05
0.32
0.38
0.03
0.22
0.25
0.11
<0.01
b
a c[Photosensitizer] = 1.0 × 10−5 M. ΦUC in deoxygenated dichlor-
oethane, where c[DPA] = 1.5 × 10−2 M, c[perylene] = 2.0 × 10−3 M, and c
c
[BPEA] = 3.3 × 10−3 M. ΦUC in aerated DMSO, where c[DPA] = 5.0 ×
10−3 M, c[perylene] = 5.0 × 10−3 M, and c[BPEA] = 6.7 × 10−4 M.
d Upconverted fluorescence was unobserved.
Table 3 lists the measured ΦUC values at the power density
beyond Ith. Although Os-diphen had the most intense absorp-
tion ability in the region, its inefficient TTET led to the slightly
poorer behavior. Os-dipy showed the worst ΦUC value due to its
shortest triplet lifetime (Table 2). Os-phen showed the best
performance among the three complexes, e.g. ΦUC = 5.9%,
2.3%, and 12.5% for DPA, perylene, and BPEA in dichloro-
Fig. 5 Effect of concentration of (a) DPA, (b) perylene, and (c) BPEA on
TTA upconversion fluorescence intensity, together with the UV-Vis
absorption and fluorescence emission spectra of DPA itself (d).
c
[Photosensitizer] = 1 × 10−5 M, dichloroethane as solvent, λex = 663 nm,
with a power density of 3800 mW cm−2, 25 °C.
(Fig. 5b), where the maximal fluorescence intensity was ethane, respectively. It is worth noting that in the TTA upcon-
obtained at 2.0 × 10−3 M concentration and then decreased version system of Os-phen/DPA, the anti-Stokes shift was up to
with increasing concentration. It is suggested that the self- 1.14 eV. This is the largest reported value of all reported solu-
absorption would be considerable for perylene at high concen- tion upconversion systems,13,27,29,39,40 and its Ith value is even
tration. In the present experiments, the optimized acceptor lower than those of similar systems.29,30 The observations
concentrations of DPA, perylene, and BPEA were determined strongly suggest that Os-phen/DPA in deoxygenated dichloro-
as 1.5 × 10−2 M, 2.0 × 10−3 M, and 3.3 × 10−3 M in deoxyge- ethane is a suitable candidate for TTA upconversion in solar
nated dichloroethane, and 5.0 × 10−3 M, 5.0 × 10−3 M, and energy applications.
6.7 × 10−4 M in DMSO, respectively.
TTA upconversion fluorescence spectra in aerated solutions
It is well known that the upconverted fluorescence intensity
has a quadratic dependence on the excitation power at low
intensities and transforms to a linear dependence at higher
power densities. Thus, an important parameter, Ith, is defined
as the threshold excitation power density of TTA upconversion.
When the power density is lower than Ith, ΦUC is linearly
dependent on the incident laser power. Once the power
density becomes greater than Ith, ΦUC reaches its maximum,
and the TTA process becomes the dominant deactivation
pathway for the triplet acceptors. Therefore, the ΦUC values at
the power density beyond Ith are usually used to evaluate the
quality of the photosensitizer.
As shown in Fig. 4d, the Ith value was determined as
1480 mW cm−2 for Os-phen with the DPA concentration of
2.0 × 10−4 M. When the DPA concentration was increased to
1.5 × 10−2 M, Ith was reduced to 200 mW cm−2. This relation-
ship agreed with the fact that the TTET efficiency was reduced
with a decrease in the acceptor concentration. According to
previous study,38 the Ith was proportional to the reciprocal of
TTET efficiency ΦTTET. In the present experiments, the linear
relationship between Ith and 1/ΦTTET was observed (ESI
Fig. S15†). For perylene and BPEA as acceptors, similar depen-
dencies of fluorescence intensity on power density were
observed (Fig. 4e and f). The Ith value was determined as
185 mW cm−2 for perylene (2.0 × 10−3 M) and 150 mW cm−2
for BPEA (3.3 × 10−3 M).
Because oxygen molecules can quickly quench triplet photo-
sensitizers, the majority of TTA upconversion systems have
been established in deoxygenated solutions. This shortcoming
significantly restricts their practical application. In order to
eliminate the effect of oxygen molecules, several interesting
methods have been developed, such as addition of oxygen
scavengers in solutions,41,42 using solutions with low dissolved
oxygen-like oils,43 or applying polyphosphates as a protective
matrix.44 However, adding such substances may affect the
absorption of the photosensitizer and fluorescence emission of
the acceptor owing to change in solvent polarity, making the
overall kinetics more complicated and reducing the ΦUC value.
Recently, a long-lived phosphorescence emission of Au com-
plexes was observed in aerated DMSO solution,45 suggesting
that DMSO is a solvent with low solubility of oxygen and
especially suitable for TTA upconversion without deoxygenation.
It was confirmed later that DMSO reacts with singlet oxygen to
produce methyl sulfone.46 Therefore, both sulfoxides and cyclic
ureas were deemed to be photochemically deoxygenating
solvents.46 Herein, the present TTA upconversion systems of
Os(II) complexes were tested in aerated DMSO solutions to
compare with those in deoxygenated dichloroethane.
The ΦUC values measured in aerated DMSO solutions are
listed in Table 3. Compared to those in deoxygenated dichloro-
ethane, the ΦUC values were much lower in aerated DMSO, e.g.,
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
0.11% for Os-phen/DPA. No upconverted fluorescence was principle, the reaction between DMSO and singlet oxygen can
observed in the case of Os-dipy/DPA. These results are unsur- quickly consume the dissolved oxygen, reducing the quench-
prising, because the fluorescence quantum yields of the accep- ing of the triplet acceptors by oxygen, thereby enhancing fluo-
tors in DMSO were clearly reduced as a result of the increase rescence. However, at a higher DMSO fraction than 30%, the
of polarity.18 Moreover, the viscosity of DMSO, ηDMSO, is as TTET efficiency was remarkably confined because of the
large as 1.996 mPa s, so that the formation rate of encounter increase of solvent viscosity. The decreased fluorescence
by collision is reduced, and confines the TTET process quantum yields of the acceptors further reduced ΦUC in the
remarkably.18
mixed solvents.
To obtain higher ΦUC in aerated solutions, a simple strategy
Surprisingly, when DPA was used as an acceptor, a larger
is to use mixed solvents of DMSO and dichloroethane, since molar fraction of DMSO led to weaker observed fluorescence.
these have reduced viscosities (ηdichloroethane = 0.84 mPa s) and This behavior is completely different from that of the other
can improve the fluorescence quantum yield of an acceptor. acceptors. Recently, it has been confirmed that DPA can react
Furthermore, the advantageous attribute of DMSO of consum- with singlet oxygen to produce endoperoxides.47 Thus, after
ing singlet oxygen can be retained as well. Fig. 6a–c show the irradiation for several seconds, the dissolved oxygen would be
TTA upconversion fluorescence spectra of Os-phen/acceptors exhausted, and then the upconverted fluorescence intensity
systems in mixed solvents with different molar fractions of would decrease monotonically with the increasing fraction of
DMSO. The dependence of the upconverted fluorescence DMSO, owing to reduced TTET efficiency and fluorescence
intensity on the DMSO molar fraction is plotted in Fig. 6d.
quantum yield of the acceptor. A similar phenomenon was
Although the phosphorescence spectra of the photosensiti- observed in the PtOEP/DPA system with air-saturated chloro-
zers were visibly changed with the varying DMSO fraction, TTA form as solvent.48 All observations show that a mixed solvent
upconversion spectra were not significantly affected except for is feasible but inefficient for achieving TTA upconversion in
their intensity (Fig. 6a–c). Interestingly, the upconverted fluo- aerated solutions.
rescence intensities of different acceptors exhibited different
dependencies on the DMSO fraction. When perylene and BPEA
were used as triplet acceptors, the upconverted fluorescence
was enhanced with the increasing DMSO fraction and reached
Conclusions
the maximum at 30% (Fig. 6d), where ΦUC was enhanced Using different ligands, three Os(II) complexes were syn-
nearly three-fold for perylene (0.8%) and BPEA (0.7%). thesized and applied as triplet photosensitizers in TTA upcon-
However, when the DMSO fraction was increased beyond 30%, version. Owing to the heavy-atom effect, direct S–T transition
the upconverted fluorescence intensity was gradually reduced. of the photosensitizers was feasible. For Os-dipy, Os-phen, and
According to approximate dissolved oxygen levels of DMSO Os-diphen, the lifetimes of the lowest triplet states were 107
and dichloroethane, the oxygen concentration itself in solu- ns, 373 ns, and 386 ns in deoxygenated dichloroethane, and
tions was not significantly changed with the DMSO fraction. In the triplet energies determined from steady-state phosphor-
escence emission spectra were 1.75 eV, 1.80 eV, and 1.74 eV,
respectively.
To reduce energy loss of TTET, DPA, perylene, and BPEA
were chosen as the triplet acceptors as their triplet energies
were relatively close. The strong upconverted fluorescence
from these triplet acceptors was observed in the visible to
violet range using the Os(II) complexes as photosensitizers.
The upconverted fluorescence quantum yields were deter-
mined as 5.9%, 2.3%, and 12.5% for Os-phen/DPA, Os-phen/
perylene, and Os-phen/BPEA, respectively, in deoxygenated
dichloroethane. It is worth mentioning that an anti-Stokes
shift of 1.14 eV was achieved from near-infrared to violet in the
Os-phen/DPA system in solutions, which is the largest shift in
all the reported experimental values to date. All the con-
clusions indicate that this system is a suitable candidate for
solar energy applications.
The performances of these TTA upconversion systems were
Fig. 6 TTA upconversion fluorescence spectra in air with Os-phen as also assessed in aerated solutions. In a typical deoxygenating
photosensitizer (1 × 10−5 M). (a) c[DPA] = 5 × 10−3 M. (b) c[perylene] = 5
solvent, DMSO, ΦUC values were significantly decreased
because of the reduced TTET efficiency and fluorescence
quantum yields, although quenching by oxygen was confined.
× 10−3 M, (c) c[BPEA] = 6.7 × 10−4 M. Mixed solvents of dichloroethane
(EDC) and DMSO were used. (d) Upconverted fluorescence intensity
versus the mole fraction of DMSO in the mixed solvents, as well as the
For comparison, mixtures of dichloroethane and DMSO were
also utilized as solvents for these TTA upconversion systems. It
photographs of upconverted fluorescence. λex = 663 nm, P = 3800 mW
cm−2, 25 °C.
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.

View Article Online
Paper
Dalton Transactions
was found that the upconverted fluorescence intensities of the
different acceptors exhibited different dependence on the
DMSO fraction. The maximal TTA upconversion quantum
yields were measured to be 4.5%, 0.8%, and 0.7% for Os-phen/
DPA, Os-phen/perylene, and Os-phen/BPEA, respectively.
Thus, change of ligands in these Os(II) complexes not only
broadened the reported anti-Stokes shift, but also promoted
the TTA upconversion quantum yield. The present work pro-
7 B. Wang, B. Sun, X. Wang, C. Ye, P. Ding, Z. Liang, Z. Chen,
X. Tao and L. Wu, J. Phys. Chem. C, 2014, 118, 1417–1425.
8 A. Monguzzi, A. Oertel, D. Braga, A. Riedinger, D. K. Kim,
P. N. Knüsel, A. Bianchi, M. Mauri, R. Simonutti and
D. J. Norris, ACS Appl. Mater. Interfaces, 2017, 9, 40180–
40186.
9 H.-I. Kim, S. Weon, H. Kang, A. L. Hagstrom, O. S. Kwon,
Y.-S. Lee, W. Choi and J.-H. Kim, Environ. Sci. Technol.,
2016, 50, 11184–11192.
vides
a
synthetic tendency of Os(II)-based triplet
photosensitizers.
10 Q. Liu, T. Yang, W. Feng and F. Li, J. Am. Chem. Soc., 2012,
134, 5390–5397.
11 A. Nagai, J. B. Miller, P. Kos, S. Elkassih, H. Xiong and
D. J. Siegwart, ACS Biomater. Sci. Eng., 2016, 1, 1206–1210.
12 J. Zhao, W. Wu, J. Sun and S. Guo, Chem. Soc. Rev., 2013,
42, 5323–5351.
13 N. Yanai and N. Kimizuka, Acc. Chem. Res., 2017, 50, 2487–
2495.
Experimental
DPA, perylene, BPEA, and methylene blue were purchased
from Aladdin Reagent (Shanghai) Co., Ltd and used without
any further purification. The synthetic procedure and struc-
tural characterization of the photosensitizers are described in
the ESI.† Further details of the steady-state spectra, transient
spectra, TTA upconversion spectra, and the corresponding
kinetic measurements are also given in the ESI.†
14 J. Peng, X. Guo, X. Jiang, D. Zhao and Y. Ma, Chem. Sci.,
2016, 7, 1233–1237.
15 H. Kouno, T. Ogawa, S. Amemori, P. Mahato, N. Yanai and
N. Kimizuka, Chem. Sci., 2016, 7, 5224–5229.
16 S. Guo, L. Ma, J. Zhao, B. Küçüköz, A. Karatay, M. Hayvali,
H. G. Yaglioglu and A. Elmali, Chem. Sci., 2014, 5, 489–500.
17 J. Wang, Y. Lu, W. McCarthy, R. Conway-Kenny,
B. Twamley, J. Zhao and S. M. Draper, Chem. Commun.,
2018, 54, 1073–1076.
Conflicts of interest
There are no conflicts to declare.
18 Q. Zhou, M. Zhou, Y. Wei, X. Zhou, S. Liu, S. Zhang and
B. Zhang, Phys. Chem. Chem. Phys., 2017, 19, 1516–1525.
19 T. N. Singhrachford and F. N. Castellano, J. Phys. Chem. A,
2009, 113, 5912.
20 Y. Wei, M. Zheng, Q. Zhou, X. Zhou and S. Liu, Org. Biomol.
Chem., 2018, 16, 5598–5608.
21 Y. Wei, M. Zhou, Q. Zhou, X. Zhou, S. Liu, S. Zhang and
B. Zhang, Phys. Chem. Chem. Phys., 2017, 19, 22049–22060.
22 W. Wu, J. Zhao, J. Sun and S. Guo, J. Org. Chem., 2012, 77,
5305–5312.
23 K. Moor, J.-H. Kim, S. Snow and J.-H. Kim, Chem.
Commun., 2013, 49, 10829.
Acknowledgements
This work was supported by the National Natural Science
Foundation of China (Grant No. 21873089, 21573208 and
21573210) and the National Key Basic Research Foundation of
China (Grant No. 2013CB834602). L. Chen is also grateful for
the financial support of the Educational Commission of Anhui
Province of China (Grant No. KJ2018A0491). DFT calculations
were performed on the supercomputing system in the
Supercomputing Center of the University of Science and
Technology of China.
24 R. Zhang, Y. Yang, S. Yang, V. S. P. K. Neti, H. Sepehrpour,
P. J. Stang and K. Han, J. Phys. Chem. C, 2017, 121, 14975–
14980.
References
25 Z. Wang and J. Zhao, Org. Lett., 2017, 19, 4492–4495.
1 C. Ye, L. Zhou, X. Wang and Z. Liang, Phys. Chem. Chem. 26 D. Wei, F. Ni, Z. Zhu, Y. Zou and C. Yang, J. Mater. Chem. C,
Phys., 2016, 18, 10818–10835. 2017, 5, 12674–12677.
2 A. Haefele, J. Blumhoff, R. S. Khnayzer and 27 J. Han, Y. Jiang, A. Obolda, P. Duan, F. Li and M. Liu,
F. N. Castellano, J. Phys. Chem. Lett., 2012, 3, 299–303. J. Phys. Chem. Lett., 2017, 8, 5865–5870.
3 A. Monguzzi, R. Tubino, S. Hoseinkhani, M. Campione and 28 D. Liu, Y. Zhao, Z. Wang, K. Xu and J. Zhao, Dalton Trans.,
F. Meinardi, Phys. Chem. Chem. Phys., 2012, 14, 4322–4332. 2018, 47, 8619–8628.
4 S. Baluschev, T. Miteva, V. Yakutkin, G. Nelles, A. Yasuda 29 Y. Sasaki, S. Amemori, H. Kouno, N. Yanai and
and G. Wegner, Phys. Rev. Lett., 2006, 97, 143903. N. Kimizuka, J. Mater. Chem. C, 2017, 5, 5063–5067.
5 Y. Y. Cheng, B. Fückel, R. W. MacQueen, T. Khoury, 30 S. Amemori, Y. Sasaki, N. Yanai and N. Kimizuka, J. Am.
R. G. Clady, T. F. Schulze, N. J. Ekins-Daukes, Chem. Soc., 2016, 138, 8702–8705.
M. J. Crossley, B. Stannowski, K. Lips and T. W. Schmidt, 31 H. Uoyama, K. Goushi, K. Shizu, H. Nomura and C. Adachi,
Energy Environ. Sci., 2012, 5, 6953. Nature, 2012, 492, 234–238.
6 S. P. Hill and K. Hanson, J. Am. Chem. Soc., 2017, 139, 32 W. L. Tsai, M. H. Huang, W. K. Lee, Y. J. Hsu, K. C. Pan,
10988–10991.
Y. H. Huang, H. C. Ting, M. Sarma, Y. Y. Ho, H. C. Hu,
Dalton Trans.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
C. C. Chen, M. T. Lee, K. T. Wong and C. C. Wu, Chem. 40 F. Deng, J. Blumhoff and F. N. Castellano, J. Phys. Chem. A,
Commun., 2015, 51, 13662–13665. 2013, 117, 4412–4419.
33 S. Altobello, R. Argazzi, S. Caramori, C. Contado, S. Da Fré, 41 C. Mongin, J. H. Golden and F. N. Castellano, ACS Appl.
P. Rubino, C. Choné, G. Larramona and C. A. Bignozzi,
J. Am. Chem. Soc., 2005, 127, 15342–15343.
Mater. Interfaces, 2016, 8, 24038–24048.
42 D. Dzebo, K. Moth-Poulsen and B. Albinsson, Photochem.
Photobiol. Sci., 2017, 16, 1327–1334.
43 C. Ye, B. Wang, R. Hao, X. Wang, P. Ding, X. Tao, Z. Chen,
Z. Liang and Y. Zhou, J. Mater. Chem. C, 2014, 2, 8507–
8514.
44 F. Marsico, A. Turshatov, R. Peköz, Y. Avlasevich,
M. Wagner, K. Weber, D. Donadio, K. Landfester,
S. Baluschev and F. R. Wurm, J. Am. Chem. Soc., 2014, 136,
11057–11064.
34 T. Kinoshita, J.-I. Fujisawa, J. Nakazaki, S. Uchida, T. Kubo
and H. Segawa, J. Phys. Chem. Lett., 2012, 3, 394–398.
35 V. Gray, D. Dzebo, M. Abrahamsson, B. Albinsson and
K. Moth-Poulsen, Phys. Chem. Chem. Phys., 2014, 16,
10345–10352.
36 S. R. Johnson, T. D. Westmoreland, J. V. Caspar,
K. R. Barqawi and T. J. Meyer, Inorg. Chem., 1988, 27, 3195–
3200.
37 Y. T. Wang, X. Y. Liu, G. Cui, W. H. Fang and W. Thiel, 45 S. Wan and W. Lu, Angew. Chem., Int. Ed., 2017, 56, 1784–
Angew. Chem., Int. Ed., 2016, 55, 14009–14013. 1788.
38 A. Monguzzi, J. Mezyk, F. Scotognella, R. Tubino and 46 S. Wan, J. Lin, H. Su, J. Dai and W. Lu, Chem. Commun.,
F. Meinardi, Phys. Rev. B: Condens. Matter Mater. Phys.,
2008, 78, 195112.
2018, 54, 3907–3910.
47 W. Fudickar and T. Linker, Chem. – Eur. J., 2011, 17, 13661–
39 Y. Y. Cheng, B. Fuckel, T. Khoury, R. G. Clady, N. J. Ekins-
13664.
Daukes, M. J. Crossley and T. W. Schmidt, J. Phys. Chem. A, 48 T. Ogawa, N. Yanai, A. Monguzzi and N. Kimizuka, Sci.
2011, 115, 1047–1053.
Rep., 2015, 5, 10882.
This journal is © The Royal Society of Chemistry 2019
Dalton Trans.