Anthraquinone-based intramolecular charge-transfer compounds: Computational molecular design, thermally activated delayed fluorescence, and highly efficient red electroluminescence

Article abstract of DOI:10.1021/ja510144h

Red fluorescent molecules suffer from large, non-radiative internal conversion rates (k_IC) governed by the energy gap law. To design efficient red thermally activated delayed fluorescence (TADF) emitters for organic light-emitting diodes (OLEDs), a large fluorescence rate (k_F) as well as a small energy difference between the lowest singlet and triplet excited states (ΔE_ST) is necessary. Herein, we demonstrated that increasing the distance between donor (D) and acceptor (A) in intramolecular-charge-transfer molecules is a promising strategy for simultaneously achieving small ΔE_ST and large k_F. Four D-Ph-A-Ph-D-type molecules with an anthraquinone acceptor, phenyl (Ph) bridge, and various donors were designed, synthesized, and compared with corresponding D-A-D-type molecules. Yellow to red TADF was observed from all of them. The k_F and ΔE_ST values determined from the measurements of quantum yield and lifetime of the fluorescence and TADF components are in good agreement with those predicted by corrected time-dependent density functional theory and are approximatively proportional to the square of the cosine of the theoretical twisting angles between each subunit. However, the introduction of a Ph-bridge was found to enhance k_F without increasing ΔE_ST. Molecular simulation revealed a twisting and stretching motion of the N-C bond in the D-A-type molecules, which is thought to lower ΔE_ST and k_F but raise k_IC, that was experimentally confirmed in both solution and doped film. OLEDs containing D-Ph-A-Ph-D-type molecules with diphenylamine and bis(4-biphenyl)amine donors demonstrated maximum external quantum efficiencies of 12.5% and 9.0% with emission peaks at 624 and 637 nm, respectively.

Full text of DOI:10.1021/ja510144h

Subscriber access provided by UNIV OF HOUSTON
Article
Anthraquinone-Based Intramolecular-Charge-Transfer Compounds:
Computational Molecular Design, Thermally Activated Delayed
Fluorescence, and Highly-Efficient Red Electroluminescence
Qisheng Zhang, Hirokazu Kuwabara, William J. Potscavage, Shuping
Huang, Yasuhiro Hatae, Takumi Shibata, and Chihaya Adachi
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/ja510144h • Publication Date (Web): 03 Dec 2014
Downloaded from http://pubs.acs.org on December 8, 2014
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Anthraquinone-Based Intramolecular-Charge-Transfer Compounds:
Computational Molecular Design, Thermally Activated Delayed
Fluorescence, and Highly-Efficient Red Electroluminescence
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
§
,†
§,†,‡
†,#
†
Qisheng Zhang, Hirokazu Kuwabara,
William J. Potscavage, Jr., Shuping
†
†
,†
Huang, Yasuhiro Hatae, Takumi Shibata, Chihaya Adachi*
†
[
] Center for Organic Photonics and Electronics Research (OPERA), Kyushu University, 744
Motooka, Nishiꢀku, Fukuoka 819ꢀ0395, Japan.
‡
[
] Research & Development Group (R&D Planning Division), Nippon Kayaku Company Ltd.,
3
ꢀ31ꢀ12 Shimo, Kita, Tokyo 115ꢀ8588, Japan.
#
[ ] Advanced Technology R&D Department, Research and Development Division, Japan
Display Inc., 3300, Hayano, Mobara, Chiba, 297ꢀ8622, Japan.
[
[
*] Eꢀmail: adachi@cstf.kyushuꢀu.ac.jp; Fax: +81ꢀ92ꢀ802ꢀ6921; Tel: +81ꢀ92ꢀ802ꢀ6920
§
] These authors contributed equally.
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ABSTRACT: Red fluorescent molecules suffer from large nonꢀradiative internal conversion rates
(
k
IC) governed by the energy gap law. To design efficient red thermally activated delayed
fluorescence (TADF) emitters for organic lightꢀemitting diodes (OLEDs), a large fluorescence rate
(k ) as well as a small energy difference between the lowest singlet and triplet excited states (ꢁEST
is necessary. Herein, we demonstrated that increasing the distance between donor (D) and acceptor
A) in intramolecularꢀchargeꢀtransfer molecules is a promising strategy for simultaneously
achieving small ꢁEST and large k . Four DꢀPhꢀAꢀPhꢀD type molecules with an anthraquinone
acceptor, phenyl (Ph) bridge, and various donors were designed, synthesized, and compared with
corresponding DꢀAꢀD type molecules. Yellow to red TADF was observed from all of them. The k
F
)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
F
F
and ꢁEST values determined from the measurements of quantum yield and lifetime of the
fluorescence and TADF components are in good agreement with those predicted by corrected
timeꢀdependent density functional theory and are approximatively proportional to the square of the
cosine of the theoretical twisting angles between each subunit. However the introduction of a
Phꢀbridge was found to enhance k
twisting and stretching motion of the NꢀC bond in the DꢀA type molecules, which is thought to
lower ꢁEST and k but raise kIC, that was experimentally confirmed in both solution and doped film.
F
without increasing ꢁEST. Molecular simulation revealed a
F
OLEDs containing DꢀPhꢀAꢀPhꢀD type molecules with diphenylamine and bis(4ꢀbiphenyl)amine
donors demonstrated maximum external quantum efficiencies of 12.5% and 9.0% with emission
peaks at 624 nm and 637 nm, respectively.
2
ACS Paragon Plus Environment

Page 3 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
■
INTRODUCTION
When the energy difference between the lowest singlet (S ) and triplet (T ) excited states (ꢁE ) is
1
1
ST
small, molecules can upconvert from the T
1 1
to S by absorbing environmental thermal energy and
then radiatively decay from the S (Scheme 1). This light emission is called thermally activated
1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
delayed fluorescence (TADF). Although TADF from many types of pure organic compounds has
1
been reported since 1930, high efficiency and short lifetime TADF emitters for application to
2
organic lightꢀemitting diodes (OLEDs) were only developed in recent years. TADF emitters in
OLEDs can convert triplet excitons, which constitute 75% of electrically generated excitons, into
light with a theoretical yield up to 100%, which could only be realized by employing
3,4
nobleꢀmetalꢀbased phosphors and Cu(I) complexes until two years ago. Although many efficient
metalꢀtoꢀligand chargeꢀtransfer (MLCT) Cu(I) complexes also show thermally activated emission
4
b,4e,5
at room temperature (RT),
they are more likely an intermediate between phosphors and
1
3
TADF emitters because of the mixing of MLCT and MLCT states by their heavy metal
4d,4e
centers.
Since heavyꢀatomꢀfree TADF emitters are cheaper than nobleꢀmetal complexes and
offer more design freedom than fourꢀcoordinate tetrahedral Cu(I) complexes, they have become a
hot topic in the fields of electroluminescence (EL) and electrochemiluminescence (ECL)
6
ꢀ9
following in the footsteps of phosphorescent materials.
Scheme 1. Jablonski diagram for a typical TADF emitter. F, fluorescence; P, phosphorescence; IC, internal
conversion; ISC, intersystem crossing. ISC represents a process from S to T in this paper unless otherwise stated.
1
1
For TADF emitters used in OLEDs, a high photoluminescence (PL) quantum yield (Φ) can
enhance the upper limit of the internal EL quantum efficiency (IQE), while a short TADF lifetime
can reduce the device efficiency rollꢀoff caused by tripletꢀtriplet annihilation (TTA), singletꢀtriplet
10
annihilation (STA), and tripletꢀpolaron annihilation (TPA). To achieve a short natural lifetime of
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
TADF, the rate constant of fluorescence (k
F
) should be large while ꢁEST should be small, with
2e
emphasis placed on the latter. On the other hand, to achieve a high Φ, k should be significantly
F
higher than the rate constant (kIC) of internal conversion (IC) from S
be small so that the rate constant of intersystem crossing (ISC) from T
higher than that from T to S . As we known, two conditions must be met for a small ꢁEST. The
1
to S
0
, and ꢁEST should also
1
to S
1
can be considerably
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
0
first is a small exchange integral (Jif = 1/2ꢁE ) for the S transition, which is roughly
ST
1
proportional to the orbital overlap integral (Sif = <
φ
i

f i f
φ >) between the initial (φ ) and final (φ )
state orbitals and corresponds to the degree of spatial overlap of the orbitals (page 63 in Ref. 11).
2b,2e
The second is that the orbitals involved in the S and T transitions are not extremely different.
1
1
A successful approach to realizing high efficiency and short lifetime TADF is to create a bipolar
molecule with a large dihedral angle (θ) between a Nꢀdonor and a phenyl ring substituted with
2
c,2e,8
electronꢀwithdrawing groups (Scheme 2, aꢀtype),
which is similar to the soꢀcalled pretwisted
1
c,12
intramolecularꢀchargeꢀtransfer (ICT) molecules.
In comparison with exciplex systems and
6
,7
quaternary carbonꢀlinked donorꢀbridgeꢀacceptor (DꢀBꢀA) systems, pretwistedꢀICTꢀtype TADF
emitters have higher k and Φ values owing to the partial overlap of orbitals on the phenyl ring
F
2
13
and an approximately proportional relationship between k
F
and Sif (see Eq. 1). ICT emitters with
2c,2e
large dihedral angles have realized efficient blue and green TADF.
However, the yellow and
orange emitters show relatively low Φ because their k
F
are not high enough to compete with kIC
2
c,2e,8
that are increasing based on the energy gap law.
Owing to the naturally small k of MLCT
F
complexes, improving Φ of red Cu(I) complexes has also been a challenge for chemists in the past
1
4
three decades.
Scheme 2. Models for donorꢀacceptor and donorꢀbridgeꢀacceptor molecules.
4
ACS Paragon Plus Environment

Page 5 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
An important design consideration for efficient red TADF is to obtain a large k as well as an
F
acceptable ꢁEST. The introduction of an additional aromatic bridge to increase the D/A separation
distance may be a solution (Scheme 2, bꢀtype). For chargeꢀtransfer (CT) transitions with weak
V
electronic coupling, the vertical transition rate (k ) can be described by an approximate expression
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
3
based on first principles as follows:
2
if
2
k
V
V
= KE H
r
(1)
where K is a constant, E
V
is the transition energy, r is the effective D/A separation distance, and
H is the electronic coupling (also called transfer integral) which is predicted to be proportional to
if
15
1
3
11
S
V
if in this case. Unlike for CT/ CT splitting, which depends mostly on Sif, k is related to more
factors including distance r. Therefore, if two ICT molecules have similar degrees of orbital
overlap (or ꢁE ), the molecule with larger r will have a larger k .
ST
F
On the basis of the previous considerations, we designed a series of symmetric DꢀπꢀAꢀπꢀD
molecules with anthraquinone (AQ) as acceptor, diphenylamine (DPA), bis(4ꢀbiphenyl)amine
(
BBPA), 3,6ꢀdiꢀtertꢀbutylcarbazole (DTC) and 9,9ꢀdimethylꢀ9,10ꢀdihydroacridine (DMAC) as
donors, and phenyl (Ph) rings as πꢀbridges (Scheme 3). The selection of AQ is due to its high T
1
1
6
energy level (2.73 eV) as well as strong electronꢀwithdrawing capability, which is another key to
realizing red TADF. For comparison, DꢀAꢀD type molecules were also synthesized by directly
linking the donor and acceptor units. In this paper, the influence of the D/A separation distance
and the dihedral angles between subunits on the photophysical properties of these AQꢀbased ICT
compounds is thoroughly investigated in terms of ꢁE , k and k .
ST
F
IC
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 3. Molecular structures of compounds a1ꢀ4 and b1ꢀ4.
■
RESULTS AND DISCUSSION
Molecular Simulations. Density functional theory (DFT) and timeꢀdependent density funcꢀ
17
tional theory (TDꢀDFT) were employed to reproduce the S and S geometries, respectively, of
0
1
these AQꢀbased ICT molecules. It is known that the configuration interaction (CI) description of
the S transition as well as the optimized geometry of the S state for an ICT molecule depend on
1
1
1
8
the exchangeꢀcorrelation (XC) functional of TDꢀDFT. The reason for this is that a CT state is
more sensitive to the HartreeꢀFock percentage (HF%) in traditional XC functionals than a
19,20
1
locallyꢀexcited state (LE),
1
and therefore the gap and coupling between a CT state (S ) and
1
higher lying LE states can be affected by using different XC functionals. Given that there is not
much difference in CT amount (q) between the S transitions based on the S and S geometries
1
0
1
for these ICT molecules, optimizing S
1
geometries on the basis of a functional with a HF% close
to
Accordingly, the S geometries of all studied molecules
to the optimal HF% (OHF) employed for the calculation of vertical absorption energy from S
0
1
8b,20
S (E (S )) may minimize the error.
1
VA
1
0
2
1
were first optimized using the B3LYP functional with 6ꢀ31G(d) basis set in vacuum as an initial
2
2
guess. The q was analyzed by Multiwfn, and the OHF for the EVA(S
1
) calculation was obtained
20
by an empirical relationship of OHF = 42q (Table S1). Then, the geometries on the S
1
state were
optimized by TDꢀDFT with a functional with HF% close to OHF at the 6ꢀ31G(d) level in toluene.
Although S geometries are generally insensitive to the XC functional and solvent, to have a better
0
comparison, their simulation was redone in toluene by using the same methods for the
corresponding S geometries.
1
6
ACS Paragon Plus Environment

Page 7 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
In the S
0
state, the twisting of DꢀπꢀA units in b1ꢀ3 can adopt two models (Scheme 4). Model I is
that the dihedral angle between D and A units is equal to that between the D unit and Phꢀbridge (θ
plus that between the A unit and Phꢀbridge (θ ). Model II is that the dihedral angle between D and
A units equals θ minus θ . The S geometry simulation of b1 reveals the same energies for these
two models, indicating an equal chance of their existence (Fig. S1). Here, we select model I for all
of the following simulations. The geometries and frontier orbitals of the S state for all eight
1
)
2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
2
0
0
molecules are shown in Fig. 1, and the important geometrical parameters are listed in Table 1. The
molecules based on DPA and BBPA donors have relatively small dihedral angles of 29ꢀ38º
between the donors and either the AQꢀacceptor (a-type) or the Phꢀbridge (b-type), θ or θ
1
. In
contrast, the molecules based on the donor DMAC have extremely large dihedral angles of 85ꢀ90º
2
e
1
due to the large steric hindrance. The dihedral angles of θ and θ in the DTCꢀbased molecules
(46ꢀ50º) are between those of the other molecules. Note that the HOMO→LUMO transitions are
all symmetry forbidden for these symmetrical molecules, while the HOMOꢀ1→LUMO transitions
with approximate energies are allowed (Table S2). Herein, the lowꢀenergy transitions
predominated by the latter are regarded as the S transition. As shown in Fig. 1, the LUMO
1
orbitals are localized mainly on the AQꢀacceptors and slightly on the adjacent nitrogen atoms or
Phꢀbridges. In contrast, the HOMOꢀ1 orbitals are mainly distributed on the donors and only
1
partially on the adjacent Phꢀbridge or AQꢀacceptor. Increasing the dihedral angles θ and θ inhibits
the delocalization of the nitrogen lone pair to the π system of the AQꢀacceptor and Phꢀbridge,
respectively, leading to a smaller overlap of the frontier orbitals and decreased coplanarity
between Phꢀbridges and the AQꢀacceptor for bꢀtype molecules (Table 1). Owing to the enlarged
D/A separation distance, the orbital overlap of a bꢀtype molecule is smaller than that of the aꢀtype
molecule with the same donor. Similarly, the orbital overlap of a BBPAꢀdonor molecule is also
smaller than that of the corresponding DPAꢀbased molecule.
Scheme 4. Two twisting models for donorꢀbridgeꢀacceptor moieties.
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
HOMO-1
LUMO
HOMO-1
LUMO
a1
a2
b1
b2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a3
a4
b3
b4
Figure 1. The HOMOꢀ1 and LUMO of the investigated molecules in their S
0
state in toluene. a1ꢀ3 optimized at
24
25
the DFT/MPW1B95 /6ꢀ31G(d) level, while a4 and b1ꢀ4 were optimized at the DFT/BMK /6ꢀ31G(d) level.
Table 1. Computed twisting angles (θ
1
, θ
2
1 2
and θ), bond lengths (l , l and l), transition energies (EVA and EVE, OHF
1
3
values), exchange energy between CT and CT (ꢁEST(CT)), oscillator strengths (fVA and fVE, OHF values) and
transition rates (kVA and kVE) of the investigated molecules in toluene on S
0 1
and S geometries optimized by the
same methods in Fig. 1 and 2.
S
0
geometry in toluene
1
S geometry in toluene
θ
1
/θ
2
or θ
l
1
/l
2
or l
E
VA(S
1
) ꢁEST(CT)
(eV)
k
VA
θ
1
/θ
2
or θ
l
1
/l
2
or l
E
VE(S
1
) ꢁEST(CT)
(eV)
k
VE
f
VA
f
VE
7
ꢀ1
7
ꢀ1
(
º)
(Å)
(eV)
2.68
2.58
2.64
2.22
2.69
2.64
2.92
2.69
(×10 s )
23.3
(º)
47
50
57
90
(Å)
(eV)
2.09
1.99
2.19
1.78
2.27
2.25
2.49
2.27
(×10 s )
2.47
a1
a2
a3
a4
b1
b2
b3
b4
29
30
46
90
1.393
1.395
1.398
1.429
0.52
0.499
0.590
0.353
1.423
1.424
1.423
1.449
0.36
0.087
0.167
0.087
0.46
25.6
0.29
4.30
0.29
16.0
0.19
2.72
0.01
0.0001 0.003
0.01
0.0001 0.002
34/31
38/34
50/36
85/37
1.410/1.487
1.413/1.488
1.412/1.490
1.431/1.491
0.25
0.510
0.457
0.255
0.002
24.0
20.7
14.2
0.094
36/29
38/29
47/29
90/29
1.385/1.473
1.402/1.478
1.404/1.479
1.448/1.484
0.26
0.312
0.334
0.195
10.5
11.0
7.87
0.18
0.21
0.12
0.16
0.002
0.001
0.0001 0.003
1
In the relaxed S state (one electron in the HOMO and one in the LUMO), HOMO and
HOMOꢀ1 localize on the different individual donors due to symmetry breaking (Fig. 2, Table S2).
The transition between HOMO and LUMO becomes allowed and lowest. Thus, the excited state
relaxation for each molecule will be discussed based on changes on the side where the HOMO is
located. In S
1
, the dihedral angles θ in a1ꢀ3 enlarge by 10ꢀ20º along with an increase in the bond
8
ACS Paragon Plus Environment

Page 9 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
lengths (l) between Nꢀatoms and AQꢀacceptors (NꢀC bonds) relative to the corresponding values
in the S state (Table 1). Such an internal rotation is somewhat like that observed in twisted ICT
0
12a
25
(
TICT) molecules, and can be explained by the soꢀcalled minimum overlap rule. That is, the
energy of a CT state can be lowered by geometry relaxation toward a total charge separation on
the donor and acceptor. Despite a few reports of the rotational relaxation of ICT molecules with
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
6
largeꢀsize diphenylamine donors in polar solvents, such a largeꢀamplitude nuclear rearrangement
in amorphous organic semiconductors has not yet been reported and needs to be further
investigated with additional experiments. On the other hand, the behavior of bꢀtype molecules in
the excited state is more like that of the biphenylꢀbridged DꢀπꢀA molecules that have been
27
thoroughly studied by M. Maus et al. In S
1
, the delocalization of the excited electron on the π
system of the Phꢀbridge and AQꢀacceptor enhance their conjugation, leading to decreased θ
2
1
(
coplanarization) and l (bond length alternation) (Table 1). The lack of a significant change in θ
2
from S to relaxed S may be attributed to the enlarged D/A separation distance and weak donor–
0
1
acceptor electronic coupling in the Franck–Condon state.
HOMO
LUMO
HOMO
LUMO
a1
a2
b1
b2
a3
a4
b3
b4
1
Figure 2. The HOMO (one electron) and LUMO (one electron) of the investigated molecules in their S state in
toluene. a1ꢀ3 optimized at the TDꢀDFT/MPW1B95/6ꢀ31G(d) level, while a4 and b1ꢀ4 were optimized at the
TDꢀDFT/BMK/6ꢀ31G(d) level.
The EVA(S
1
) and EVE(S
1
) (vertical emission energy) for all eight molecules were calculated in
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
toluene by using the OHF/TDꢀDFT method on the basis of the optimized S and S geometries,
0
1
20
respectively. Because of the logarithmic relationship between the vertical transition energy and
2
e,20
the HF% in XC functionals (HF% ≥ 20%),
B3LYP (20% HF) and BMK (42% HF) functionals with 6ꢀ31G(d) basis set, and EVA(S
, OHF) were read from a straight line between the two transition energy–HF% points (Table
E
VA(S
1
) and EVE(S
1
) were first calculated using the
1
, OHF) and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
E
VE(S
1
S2) plotted on a log–log scale. According to the calculation, these molecules have approximate CT
absorption peaks at around 460ꢀ480 nm, except for a4 (560 nm) and b3 (425 nm), and the energy
differences of 0.4–0.6 eV between the absorption and emission maxima. In S , since the CI
1
1 1
description of the T transitions reproduced by TDꢀDFT/B3LYP is similar to that of the S
1
1
3
transitions ( CT), the transitionꢀenergyꢀrelated CTꢀ CT splitting can be calculated by the
2
0
following formula:
ꢁE
ST(CT) = EVE(S
1
, OHF)ꢀEVE(T
1
, B3LYP)×EVE(S
1
, OHF)/EVE(S
1
, B3LYP).
(2)
1
3
The exchange energy between CT and CT transitions (HOMOꢀ1→LUMO) in a S
geometry can
0
1
3
be calculated in the same way (Table 1). The E0ꢀ0( CT) and E0ꢀ0( CT) can then be evaluated by
1
E
0ꢀ0( CT) = [EVE(S
1
, OHF)+EVA(S
1
, OHF)]/2
(3)
(4)
3
1
E
0ꢀ0( CT) = E0ꢀ0( CT)ꢀꢁEST(CT)
3
where ꢁEST(CT) is computed on the basis of the S
1
geometries. On the other hand, E0ꢀ0( LE) of all
of the investigated molecules can be predicted from their E (T , B3LYP) (Table S2) by an
VA
3
20
experience coefficient method:
3
E
0ꢀ0( LE) = EVA(T
3
, B3LYP)/CꢀꢁEStokes
(5)
where C is a correction factor found to be 1.02 for the B3LYP functional and ∆EStokes is the
Stokesꢀshift energy loss for a LE transition in toluene (0.07 eV) (See the Experimental Section in
SI for more details).
3
3
Since the calculated E0ꢀ0( LE) are all higher than or close to the E0ꢀ0( CT) (Table S1), the
computed ꢁEST(CT) ranging from 0.52 to 0.001 eV can be taken to be their ꢁEST, which is seen to
depend strongly on the degree of orbital overlap (Fig. 1 and 2). At a qualitative level, if we assume
that these shortꢀrange ICT have a proportional relationship between S and H like that found in
if
if
15
longꢀrange CT and intermolecular CT, ꢁEST(CT) is expected to be related to Jif, Sif, and, in turn,
H
if. Additionally, the electronic coupling between a directly connected donor and acceptor was
10
ACS Paragon Plus Environment

Page 11 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
27e,28
found to be proportional to the cosine of their dihedral angle,
while the electronic coupling of
a donorꢀbridgeꢀacceptor molecule is proportional to the product of the electronic coupling of the
15a,28b
donor and bridge and that of the bridge and acceptor via a superexchange mechanism.
ꢁEST(CT) values of these ICT molecules are expected to be related to cosθ and cosθ
aꢀtype and bꢀtype molecules, respectively. In fact, a proportional relationship was found between
Thus,
1
cosθ for
2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
2
2
ꢁ
E
ST(CT) and cos θ or cos θ
1 2
cos θ on the basis of our calculation results (Table S3), which is
2
29
consistent with the prediction that Jif is weighted by Sif
.
Owing to the increase of the dihedral
angle θ, the ꢁEST for a1ꢀ3 in the S state are significantly smaller than those in the S state. On the
1
0
contrary, the exchange energy for b1ꢀ3 in the S
1
0
state is slightly larger than that in the S state due
2
to the decrease of the dihedral angle θ .
TDꢀDFT also provides a relative oscillator strength (f) for each transition. Since oscillator
strength of vertical absorption and emission is a function of transition energy, their optimal values
were determined by a correction of those calculated at the TDꢀDFT/B3LYP/6ꢀ31G(d) level using
f
VA(OHF) = fVA(B3LYP)×EVA(S
1
, OHF)/EVA(S , B3LYP),
1
, OHF)/EVE(S , B3LYP).
1
(6a)
(6b)
f
VE(OHF) = fVE(B3LYP)×EVE(S
1
ꢀ1
Taking a simplified relationship that the transition rate (s ) can be approximated by the product of
ꢀ1
the oscillator strength and the square of the wavenumber (cm ) (page 196 in Ref. 11), the
transition rates of vertical absorption and emission can be calculated by
2
k
VA = fVA(OHF)×EVA(S
1
, OHF)
(7a)
(7b)
2
kVE = f (OHF)×E (S , OHF)
VE
VE
1
2
Unsurprisingly, the computed kVA and kVE are approximately proportional to cos θ for aꢀtype
2
2
molecules and cos θ
1
cos θ
2
for bꢀtype molecules (Table 1 and S3) due to a roughly proportional
2
dependence of k
V
on the dihedralꢀangleꢀdependent Hif in these ICT systems (Eq. 1). Since θ for
a4 and θ for b4 are predicted to be nearly 90º in both S and S , their ICT transitions are
1
0
1
theoretically forbidden with transition rates that are nearly zero. However, the transitions may
actually be more permissible due to a conformational distribution including more planar
geometries around the equilibrium one. For the same reason, their actual ꢁE (CT) may also be
ST
larger than the simulated values.
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
In S , the k of each aꢀtype molecule is similar to that of the bꢀtype molecule with the same
0
VA
donor. For bꢀtype molecules, the kVE in brokenꢀsymmetry S
because the S
(HOMO→LUMO) and S (HOMOꢀ1→LUMO) transitions share the overall f
(Table S2). If so, the symmetrical design (e.g., DꢀAꢀD) will not lead to a higher k . For aꢀtype
molecules, the f(S ) in S are much smaller than those in S due to not only the symmetry breaking
but also the internal rotation. Interestingly, it is found that the DTCꢀbased molecule b3 in S
1 0
are nearly half of their kVA in S
1
2
F
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
1
0
0
has a
but a
2
c,2e,8
small ꢁEST (0.12 eV), which is comparable to those of the large twisted TADF emitters,
large transition rate that is comparable to the DPAꢀbased molecule b1, suggesting that a large kF
and a small ꢁEST can be achieved in one molecule by adopting a bꢀtype structure with well
1 2
controlled θ and θ .
Photophysical Properties. The synthesis of the eight compounds in Scheme 3 is described in
the supporting information. Their absorption and emission spectra in toluene at RT are shown in
1
Fig. 3. The lowestꢀenergy absorption bands (400ꢀ500 nm) can be ascribed to the CT transitions.
Their peak energies are consistent with the computed EVA(S
emissive a4 and yellowꢀemitting b3, the molecules emit orangeꢀred to red light with peaks at
00ꢀ630 nm in toluene. The computed EVE(S ) for aꢀtype molecules coincide with their emission
1
) (Table S1). Except for the hardly
6
1
peak energies, while those of bꢀtype molecules are generally overestimated by 0.2 eV (Table S1).
The polarizable continuum model used here may underestimate the solvation in
3
0
1 1
toluene/bꢀtypeꢀmolecule systems, because the EVA(S ) and EVE(S ) calculated under vacuum are
in good agreement with the experimental values in nonpolar solvent (Table S4, Fig. S2b). The
1
oscillator strength of the CT absorption listed in Table 2 were determined experimentally from the
integration of its band by (page 195 in Ref. 11)
ꢀ9
f = 4.3×10 ʃɛdṽ,
(8)
where ɛ is the molar extinction coefficient and ṽ is the wavenumber of the absorption. The f values
for phenylamine, DTC, and DMAC derivatives decrease in order, while a- and b-type molecules
with the same donor have similar f. Except for the DMAC derivatives having significantly higher f
than the computed ones, the experimental f for the molecules are about twoꢀthirds of the computed
ones (Table 1).
12
ACS Paragon Plus Environment

Page 13 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. The absorption and emission spectra of the investigated molecules in toluene (0.02 mM) at RT. The
lowꢀintensity emission spectrum of a4 is shown in Fig. S2a.
Table 2. Photophysical data for the investigated molecules in toluene at RT.
in airꢀsaturated toluene
2
in O ꢀfree toluene
λ
abs,max
f
λ
em,max
Φ
τ
F
k
F
k
SB
Φ
Φ
F
τ
F
τ
TADF
k
F
7
ꢀ1
7
ꢀ1
7 ꢀ1
(nm)
449
464
463
527
454
456
426
446
(nm)
601
629
602
662
597
608
558
609
(ns)
3.6
2.6
4.0
–
(×10 s ) (×10 s )
(ns)
5.3
3.5
7.5
–
(ꢂs)
377
74
(×10 s )
a1
a2
a3
a4
b1
b2
b3
b4
0.36
0.37
0.18
0.015
0.36
0.30
0.19
0.018
0.09
0.06
0.07
–
2.5
2.3
1.8
–
12.4
12.0
5.8
0.20
0.10
0.15
–
0.13
0.08
0.13
–
2.5
2.3
7.7
–
1.7
–
–
0.30
0.25
0.43
0.02
4.9
5.1
9.4
6.7
6.1
4.9
4.6
0.30
12.2
10.0
7.4
0.59
0.62
0.60
0.17
0.55
0.55
0.58
0.13
9.0
11.2
12.6
42
102
85
6.1
4.9
17.0
3.5
4.6
0.63
0.31
According to the Strickler–Berg relationship, given little change of the molecular structure
taking place in S , the rate constant of the emission is related to the oscillator strength of the first
absorption transition by (page 196 in Ref. 11)
1
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
kSB = 0.67ṽ f,
(9)
where ṽ is the wavenumber of the absorption maximum of the CT band. For bꢀtype molecules
measured in airꢀsaturated toluene, experimental k calculated by dividing Φ by the
singleꢀexponential fluorescence lifetime (τ ) is about half of kSB (Table 2). Taking into account the
F
F
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
simulation results, we deduce that the halving of the rate constant can be attributed to the
excitedꢀstate symmetry breaking, which can be further confirmed by comparing the kSB and
F
experimental k of an asymmetrical analogue of b1, DPAꢀPhꢀAQ (Fig. S3). However, for aꢀtype
molecules a1ꢀ3, experimental k is significantly lower than k (Table 2), which is in agreement
F
SB
with the change of transition rate between S
0
and S
1
revealed by the molecular simulations,
indicating a fast relaxation by twisting in the excited state in toluene. The experimental trends of
kSB and k determined from the absorption and emission processes, respectively, match the
F
F
computed trends of kVA and kVE but with a relationship of kSB ~ 0.5 kVA and k ~ 0.5 kVE (Table 1
and 2).
After oxygen degassing by nitrogen bubbling, Φ and τ
F
of these compounds in toluene at RT
increase significantly, accompanied by the presence of TADF with a singleꢀexponential lifetime
(τ
F
TADF) ranging from a few to hundreds of microseconds (Fig. S4). The relatively short τ in
1
airꢀsaturated toluene indicates that the dissolved oxygen can quench the longꢀlived CT states as
31
well as the triplet states of these molecules. In oxygenꢀfree toluene, the total Φ of a1ꢀ3 are
around 0.1ꢀ0.2, while those of b1ꢀ3 can be as high as nearly 0.6. Due to the lack of orbital overlap,
the Φ of b4 (0.17) is dramatically lower than that of other bꢀtype molecules. The fluorescence and
TADF components in the transient decay spectra can be distinguished by the extension of a line
fitted to the TADF decay (Fig. 4), a solution to the broad instrument response for the fluorescence
decay. The fluorescent (Φ
from the total Φ and the proportion of the integrated area of each of the components in the
transient spectra to the total integrated area. The k values calculated from Φ and in
oxygenꢀfree toluene are found to be equal to those in airꢀsaturated solutions (Table 2).
F
) and TADF (ΦTADF) components of Φ are subsequently determined
F
F
τ
F
1
4
ACS Paragon Plus Environment

Page 15 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Transient decay spectra of b1 in an O ꢀfree toluene at RT. The wide impulse response function (flash
2
lamp) causes the distortion of the curve (pink) corresponding to the prompt component.
With the goal of application in devices, the photophysical properties of the title compounds
were investigated by doping them into 4,4′ꢀbis(carbazolꢀ9ꢀyl)biphenyl (CBP), which is the most
commonly used host material for green and red phosphors and has a high E (T ) of 2.6 eV (Table
0
ꢀ0
1
3
). To lower aggregation induced redꢀshifting and quenching (see Fig. S5) so as to better compare
the data in solution and films, a low doping concentration of 1 wt% was adopted in the following
PL studies. In comparison with the emission spectra in toluene, those in 1wt%ꢀdoped CBP films
1
are blueꢀshifted by 2ꢀ62 nm (Fig. S6, Table 2 and 3) because the highly polar CT states cannot be
stabilized by the reorganization of surrounding polar molecules when doped in a solid matrix.
Except for b3, the fluorescence and TADF decay for these molecules in doped films can be well
fitted by a single or a double exponentials (Fig. 5a and S7), and the individual Φ and Φ
can
TADF
F
be calculated in the same way as in toluene. Weighted average fluorescence lifetimes (Table 3)
were used to calculate k for the molecules with double exponential fluorescence decay. For the
molecules other than b3, their k in CBP film are close to those in toluene (Table 2), implying that
F
F
aꢀtype molecules can undergo a 10ꢀ20º rotational excited state relaxation even in organic
semiconductor films. However, the excitedꢀstate conformation distribution of aꢀtype molecules
with large donor units should be wide due to their heterogeneous microenvironment in the solid
films. Not surprisingly, the streak image (300 K) of an a2ꢀdoped film reveals a blueꢀshifted
emission band based on its longꢀlived TADF component (> 2.5 ms) (Fig. 5a insert). The
component may arise from rotamers that are more planar and whose rotational relaxations are
restricted by their rigid microenvironment.
15
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. Transient decay spectra of (a) a2 and (b) b3 doped into CBP films (1 wt%) measured with a
fluorescence lifetime spectrometer (C11367ꢀ03, Hamamatsu Photonics) at RT. Black curves were recorded in
TCC900 mode; blue curves were recorded in M9003ꢀ01 mode; red curves are double exponential fitting data. Inset:
The emission spectra of TADF components measured by a streak camera at RT.
Table 3. Photophysical data for the investigated molecules doped into CBP films (1 wt%) at RT.
λ
em,max
Φ
Φ
F
τ
F
τ
TADF
k
F
k
IC
k
ISC
k
TADF
ꢀ1
ꢁEST
(eV)
0.29
0.27
0.17
0.08
0.24
0.22
–
7
ꢀ1
7
ꢀ1
7
ꢀ1
4
(
nm)
593
603
575
600
594
601
550
564
(ns)
4.9
(ꢂs)
4600
1560
(×10 s )
2.7
(×10 s )
2.6
(×10 s )
15
(×10 s )
0.011
0.027
0.84
5.1
a
a
a1
a2
a3
a4
b1
b2
b3
0.50
0.42
0.52
0.08
0.80
0.76
0.65
0.50
0.13
0.17
0.16
0.05
0.54
0.54
0.12
0.14
5.6
3.0
4.2
11
a
10.4
16.5
10.2
10.2
62
1.5
1.4
6.5
a
1.6
0.30
5.3
3.5
2.3
a
416
185
1.3
3.2
0.19
0.41
–
a
5.3
1.7
2.8
a
33
41
16/156
0.36
0.34
0.19
0.34
2.4
a
a
b4
6.5
1.7
7.7
0.07
a
2
An average lifetime calculated by
shown in Fig. S7).
i i i i i i i i
τav = ΣA τ /ΣA τ , where A is the preꢀexponential for lifetime τ (A and τ are
The photophysical properties of b3 in doped CBP films are special among these molecules. Its
16
ACS Paragon Plus Environment

Page 17 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
fluorescence decay is double exponential while its TADF decay is a multiꢀexponential that can be
roughly divided into two components (Fig. 5b). The fast component (68%) of TADF has a single
exponential decay time of 16 ꢂs, while the slow one (32%) has a single exponential decay time of
156 ꢂs. The two TADF components appear to arise from two different rotamers in the doped film
because of the different spectra observed for these two components from streak images (Fig. 5b
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
6
ꢀ1
insert). Since the k
F
value of b3 (3.3×10 s ) in the doped film is significantly smaller than that in
7
ꢀ1
toluene (4.6×10 s ), the predominant rotamer in the doped film should be more tortile than the
one in toluene. The formation of differently twisted b3 can be explained with Scheme 4. In the
0
doped film, about threeꢀfourths of b3 molecules in S contain at least one DꢀπꢀA moiety twisted
like model I (IꢀI, IꢀII, IIꢀI), while oneꢀfourth of them are only twisted like model II (IIꢀII). For the
former, the flattening distortion of the Phꢀbridge with respect to AQꢀacceptor in S may enlarge
1
the dihedral angle θ
a decrease of kF, ꢁEST, and τTADF. In contrast, the species containing only modelꢀII twisting tend to
relax toward a more planar conformation in S with increased k
1
between the Phꢀbridge and the rotationally restricted DTCꢀdonor, resulting in
1
F
, ꢁEST, and τTADF.
Since TADF emitters in rigid matrices cannot undergo deactivation through collisions with
surrounding molecules (page 288 in Ref. 11), their Φ in doped CBP films are considerably higher
than those in solution (table 2 and 3). Even a4, which showed almost no emission in toluene, has a
Φ of 0.08 in the doped film. According to the Boltzmann distribution, the time the title molecules
spend in T
1 1 1 1
is much longer than that in S when S and T achieve thermal equilibrium, so the
collisionꢀfree condition favors the suppression of nonꢀradiative decay from T to S . This is
1
0
consistent with the fact that the Φ and τF in doped films are comparable with the corresponding
F
values in oxygenꢀfree toluene while the ΦTADF and τTADF in the former are much higher than those
in the latter (Table 2 and 3). In particular, a1 and a2 show extremely long TADF lifetimes of 4.6
and 1.6 ms in the doped films (Fig. S7). Since the IC process is generally 4–8 orders of magnitude
32
faster than ISC with a similar band gap, it seems reasonable to assume that the IC process from
33
S
1
to S
0
dominates the deactivation of the TADF emitters in the doped films; therefore, the kIC
,
kISC and kTADF can be given by the following formulas.
Φ
Φ
= k
F
/(k
F
+kIC
)
(10)
(11)
F
= k
F
/(k
F
+kIC+kISC)
17
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Φ
IC = kIC/(k
ISC = 1ꢀ
F
+kIC+kISC
)
(12)
(13)
(14)
Φ
Φ
F
ꢀΦ
IC = kISC/(k +kIC+kISC)
F
kTADF
= ΦTADF/(ΦISC×τTADF)
where
approximates
CBP films seems to be governed by the DꢀA distance as well as the energy gap law (i.e., ln(k ) ∝
Φ
ISC is the triplet yield ignoring the reverse ISC from T
1
to S
1
. Note that kTADF
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
e
Φ
TADF/τTADF only in the case of kISC>>kIC+k . As shown in Fig. 6, the kIC in doped
F
IC
ꢀ1
E
VE(S
1
) ). For the same EVE(S
of a bꢀtype molecule, owing to the soꢀcalled free rotor and loose bolt effects in aꢀtype molecules
page 288 in Ref. 11). As predicted by TDꢀDFT, the S of aꢀtype molecules can be stabilized by
1
), the kIC of an aꢀtype molecule is more than twice as high as that
(
1
twisting and stretching of the NꢀC bond between a donor and acceptor, which shows some πꢀbond
character when the twisting angle θ is smaller than 90º. According to radiationless transition
1
theory, twisting a π bond or stretching a σ bond in S causes an approaching of surfaces that
facilitates IC to S (Scheme 5). The numerous investigations on MLCT Cu(I) complexes also
0
indicated that a rigid conformation is crucial to suppress radiationless relaxation of the exciplex
14
back to ground state and improve its
Φ
.
Since the free rotor and loose bolt effects can also
operate to facilitate the ISC process, the values of kISC for aꢀtype molecules are also found to be
higher than those for bꢀtype molecules (Table 3).
Figure 6. Plot of the rate constants of internal conversion between S and S (k ) versus vertical emission energy
1
0
IC
(EVE) for the investigated molecules doped into CBP films (1 wt%) at RT. The reference data is from reported
TADF emitters based on a diphenylsulphone acceptor in various solvents (SI in Ref. 2e).
18
ACS Paragon Plus Environment

Page 19 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
Scheme 5. The stretching of a σ bond or twisting of a π bond lowers the energy of CT state (S
1
) but raises the
energy of the ground state (S ), resulting in the S and S surfaces approaching each other and creating a more
0
1
0
efficient radiationless deactivation pathway (energy gap law).
Due to the low phosphorescence yield of the title molecules in toluene and doped films at low
temperatures (Fig. S8), it is hard to reliably estimate ꢁEST from their fluorescence and
phosphorescence spectra. In addition, solvent glasses can restrict the rotational relaxation of these
molecules in the excited state, and therefore ꢁEST values obtained in them do not represent the
ones at RT. Here, we calculated the ꢁEST of these molecules in doped films at RT by using an
2
e
approximate relationship among ꢁEST, kTADF, and k
F
as follows:
k
TADF = 1/3k
F
exp(ꢀꢁEST/RT)
(15)
where R and T denote the ideal gas constant and absolute temperature, respectively. Equation 15 is
simplified from a more general expression for the temperature dependence of the lifetime of two
thermally equilibrated emitting states when the gap between these two states is in the range of
2e,5b
0
.05 to 0.30 eV.
The ꢁEST of these molecules in doped films given by Eq. 15 are close to the
geometries (Table 1 and 3). On the other hand, the
experimental ꢁEST of a1ꢀ3 are significantly lower than the corresponding values simulated in the
geometries, suggesting that intramolecular twisting relaxations occur in doped films after
excitation. Since ꢁEST is insensitive to kTADF (e.g., ꢁEST decreases only 0.01 eV when kTADF
doubles), we can also estimate the ꢁEST of some efficient emitters (k > kIC) in solution using Eq.
4 and 15 with an approximation of . The reliability of such a treatment has been
confirmed by using reported TADF emitters with ꢁEST determined in toluene (Table S5).
values computed based on their S
1
S
0
F
1
Φ
ISC ≈ 1ꢀ
Φ
F
19
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Device Characterization. From the previous comparison, we can conclude that the bꢀtype
molecules are more suitable for application in OLEDs owing to their relatively high
Φ and short
τ
TADF in doped films. Consequently, the four bꢀtype molecules were selected for characterization in
34
devices with an optimized structure (Fig. S9), indium tin oxide (ITO)/HATꢀCN (10
nm)/TrisꢀPCz (30 nm)/10 wt% TADF:CBP (30 nm)/T2T (10 nm)/BpyꢀTP2 (40 nm)/LiF
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
0.8 nm)/Al, where dipyrazino[2,3ꢀf:20,30ꢀh]quinoxalineꢀ2,3,6,7,10,11ꢀhexacarbonitrile
HATꢀCN), 9,9'ꢀdiphenylꢀ6ꢀ(9ꢀphenylꢀ9Hꢀcarbazolꢀ3ꢀyl)ꢀ9H,9'Hꢀ3,3'ꢀbicarbazole TrisꢀPCz),
,4,6ꢀtris(biphenylꢀ3ꢀyl)ꢀ1,3,5ꢀtriazine T2T), 2,7ꢀbis(2,20ꢀbipyridineꢀ5ꢀyl)triphenylene
BpyꢀTP2), and LiF act as hole injection, hole transport, excitonꢀblocking, electron transport, and
(
(
2
(
(
electron injection layers, respectively. An energy diagram of the devices is shown in Fig. 7a. The
HOMO and LUMO levels of b1ꢀ4 are above and below those of the CBP host, suggesting that
holeꢀelectron capture by these emitters in the emitting layers (EMLs) can be balanced. To achieve
a complete Dexter energy transfer from the CBP host to the TADF emitters, a relatively high
doping concentration of 10 wt% was adopted for these TADF OLEDs (the
dopingꢀconcentrationꢀdependent PL and EL performance of b1 in a CBP host can be found in
Fig. S5 and S10). However, similar to most ICT molecules, these TADF emitters also suffered
from aggregationꢀinduced quenching to some degree, accompanied by an emission redꢀshift. The
emission maxima and
Φ of b1ꢀ4 in 10 wt% doped CBP films are 613 nm and 0.55, 626 nm and
0
.40, 565 nm and 0.41, and 570 nm and 0.32, respectively.
20
ACS Paragon Plus Environment

Page 21 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. (a) Energy diagram of OLEDs. HOMO levels of CBP and b1ꢀ4 were measured by photoꢀelectron
spectroscopy (ACꢀ3). Their energy gaps were determined from the absorption edge, and their LUMO levels were
calculated from the HOMO minus the energy gap. HOMO and LUMO levels of other materials are taken from
3
4
2
literature values. (b) The EL spectra at 10 mA/cm (inset: image of b1ꢀbased OLED), (c) luminanceꢀcurrent
densityꢀvoltage characteristics and (d) EQEꢀcurrent density characteristics of the OLEDs based on b1ꢀ4.
The EL spectra of these OLEDs are shown in Fig. 7b. They have a small redꢀshift (~10 nm)
with respect to the PL spectra of the EMLs. The b1ꢀ and b2ꢀbased OLEDs emit red light with
emission peaks at 624 and 637 nm and Commission Internationale de L’Eclairage (CIE)
coordinates of (0.61, 0.39) and (0.63, 0.37), respectively. On the other hand, the b3ꢀ and b4ꢀbased
OLEDs emit orange light with emission peaks at 574 and 584 nm, respectively. As shown in Fig.
2
7
c, these four devices turnꢀon at 3 V and achieved maxima luminance of 10,000ꢀ30,000 cd/m at
about 12 V. As shown in Fig 7d, the maxima external EL quantum efficiencies (EQEs) of the
devices based on b1ꢀ4 are 12.5%, 9.0%, 9.0% and 6.9%, respectively. Based on the investigations
of the PL properties in doped films (see also Ref. 33), we can conclude that S →S IC is the major
1
0
radiationless deactivation pathway for these TADF emitters in OLEDs with low doping
concentrations at low current density. As a consequence, the triplet TADF emitters will have the
same quantum yields as the singlet ones, and the maximum IQEs of the devices will be close to
21
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
the total
Φ
of the TADF dopants in EMLs even though threeꢀfourths of excitons generated by
holeꢀelectron recombination in OLEDs are in the T
1
state. Although the complicated bimolecular
quenching processes cannot be ignored at higher doping concentrations like 10 wt%, the
maximum IQEs of these TADF OLEDs are close to the
outꢀcoupling efficiency of 0.2ꢀ0.3 is assumed.
Φ of their EMLs when a light
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Although the EQEs of the red TADF OLEDs containing b1 and b2 at low current density are
comparable with those of the devices containing tris[1ꢀphenylisoquinolinatoꢀC2,N]iridium(III)
3c
(Ir(piq)
3
), a classical red phosphorescent material, their efficiencies at high current densities are
lower than those of the Ir(piq)
3
ꢀbased OLED (Fig. S11). The EQEs of b1ꢀ and b2ꢀbased OLEDs
2
decrease to 8.1% and 5.7% at a luminance of 100 cd/m and to 2.3% and 1.7% at a luminance of
2
1000 cd/m , respectively. The fast EQE rollꢀoff with increasing current density can be attributed to
the long TADF lifetimes of b1 and b2 in doped films, which are two orders of magnitude higher
3
c
than that of Ir(piq)
3
. As revealed by numerous previous studies, longꢀlived triplet excitons are
more easily be quenched by TTA, STA, and TPA processes, which are highly dependent on the
10
current density in the device. Although the EQE rollꢀoff of the b1ꢀ and b2ꢀbased OLEDs is not
observably improved over that of reported orangeꢀred TADF OLEDs (λmax = 610 nm) based on a
2d
diphenylamine/heptazine derivative (HAPꢀ3TPA), the ꢁE have been successfully reduced from
ST
0
.30 eV for HAPꢀ3TPA (by OHF/TDꢀDFT) to 0.21 eV for b2 without significantly lowering the Φ
in doped films. Since b4 in CBP has a short TADF lifetime on the order of microseconds, the EQE
2
rollꢀoff is significantly reduced in its device with an EQE at 1000 cd/m of 94% of the maximum.
In doped CBP films, a fraction of the b3 rotamers have a short TADF lifetime comparable with
that of b4 while others have a long TADF lifetime comparable with that of b2. As a result, the
degree of efficiency rollꢀoff for b3ꢀbased OLEDs is between those of b2ꢀ and b4ꢀbased OLEDs.
Despite good performance at high current density, it is worth noting that the b4-based OLED
emits orange light. If the highly twisted ICT molecules have redder emission, their
Φ will
dramatically decrease because their low k values cannot compete with the quickly growing kIC
F
,
just as in the case of b4 in toluene. Finding a balance between k and ꢁEST by adjusting the
F
twisting angles in new bꢀtype molecules is our next step but also challenge.
22
ACS Paragon Plus Environment

Page 23 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
■
CONCLUSIONS
Increasing the transition rate without enlarging the overlap of the involved orbitals is a key in the
realization of efficient red TADF emitters having fast TADF. Herein, we approached this goal by
enlarging the effective separation distance of appropriate donors and an acceptor to increase the
dipoleꢀmoment change in the CT transition. The comparison of eight TADF emitters with DPA,
BBPA, DTC, and DMAC donors and an AQ acceptor indicates that DꢀPhꢀA (bꢀ) type TADF
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
emitters have higher k
F
and consequently higher Φ relative to DꢀA (aꢀ) type TADF emitters with
the same ꢁE . In addition, it was found that the NꢀC bond linking the donor and acceptor in DꢀA
ST
type emitters twists and stretches in S
1
to stable the excited state, leading to smaller ꢁEST, lower k
F
,
and more efficient radiationless S →S
1
0
IC and S →T ISC processes. The rotational excitedꢀstate
1
1
relaxation was found to occur even in organic semiconductor films. Since most DꢀPhꢀA type
emitters show only minor geometry rearrangement in the excited state, they have smaller kIC
compared to the DꢀA type emitters with the same emission energy, which further contributes to
their higher
Φ. A DꢀPhꢀA type emitter based on DPA as donor (b1) exhibits a high Φ of 0.55 with
an emission peak at 614 nm and a relatively short TADF lifetime of 120 ꢂs in 10 wt%ꢀdoped CBP
film. The OLED using this doped film as EML achieved an EQE up to 8.1% at a luminance of 100
2
cd/m with an emission maximum at 624 nm, which are comparable to the best red PHOLEDs.
It is noteworthy that the quantum yield of red TADF emitters can be further improved by the
selection of a new πꢀbridge with longer axial length and donors and acceptors with larger atomic
orbital coefficients. Molecular design will be facilitated by using an optimal HF exchange method
based on TDꢀDFT, which has effectively reproduced various photophysical parameters of studied
TADF molecules. The molecule simulation in this work suggested that a fast transition rate and a
small ꢁEST can be achieved in one molecule by well controlling the twisting angles of a bꢀtype
molecule. However, the design idea may be partially unfulfilled due to the flattening of the
Phꢀbridge and acceptor and a subsequent twisting of the Phꢀbridge and donor in doped films.
Therefore suppressing undesirable rotational relaxation in the excited state is as important as
controlling the twisting angles between the subunits in S . In addition, suppressing rotational
0
relaxation generally results in a more narrow emission band, allowing the realization of purer red
colors without lowering the energy gap.
23
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
■
ASSOCIATED CONTENT
Supporting Information
Supplemental tables and figures; experimental section; synthesis and computational geometries of
the investigated compounds. This material is available free of charge via the Internet at
http://pubs.acs.org.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
■
AUTHOR INFORMATION
Corresponding Author
adachi@cstf.kyushuꢀu.ac.jp
■
ACKNOWLEDGMENTS
This work was supported by the Funding Program for WorldꢀLeading Innovative R&D on Science
and Technology (FIRST) and the Exploratory Research for Advanced Technology (ERATO). The
authors wish to thank Prof. Katsumi Tokumaru and Dr. Katsuyuki Shizu for stimulating
discussions with regard to this work.
■
REFERENCES
(1) (a) Boudin, M. S. J. Chim. Physique 1930, 27, 285. (b) Saltiel, J.; Curtis, H. C.; Metts, L.;
Miley, J. W.; Winterle, J.; Wrighton, M. J. Am. Chem. Soc. 1970, 92, 410. (c) Herbich, J.;
Kapturkiewicz, A.; Nowacki, J. Chem. Phys. Lett. 1996, 262, 633. (d) Heinz, B.; Schmidt, B.;
Root, C.; Satzger, H.; Milota, F.; Fierz, B.; Kiefhaber, T.; Zintha, W.; Gilch P. Phys. Chem. Chem.
Phys. 2006, 8, 3432. (e) Baleizao, C.; BerberanꢀSantos, M. N. Ann. N.Y. Acad. Sci. 2008, 1130,
224.
(
2) (a) Endo, A.; Sato, K.; Yoshimura, K.; Kai, T.; Kawada, A.; Miyazaki, H.; Adachi, C. Appl.
Phys. Lett. 2011, 98, 083302. (b) Zhang, Q.; Li, J.; Shizu, K.; Huang, S.; Hirata, S.; Miyazaki, H.;
Adachi, C. J Am. Chem. Soc. 2012, 134, 14706. (c) Uoyama, H.; Goushi, K.; Shizu, K.; Nomura,
H.; Adachi C. Nature 2012, 492, 234. (d) Li, J.; Nakagawa, T.; MacDonald, J.; Zhang, Q.; Nomura,
H.; Miyazaki, H.; Adachi, C. Adv. Mater. 2013, 25, 3319. (e) Zhang, Q.; Li, B.; Huang, S.;
Nomura, H.; Tanaka, H.; Adachi, C. Nat. Photonics 2014, 8, 326.
(3) (a) Ma, Y.; Zhang, H.; Shen, J.; Che, C. Synth. Met. 1998, 94, 245. (b) Baldo, M. A.;
O’Brien, D. F.; You, Y.; Shoustikov, A.; Sibley, S.; Thompson, M. E.; Forrest, S. R. Nature, 1998,
395, 151. (c) Tsuboyama, A.; Iwawaki, H.; Furugori, M.; Mukaide, T.; Kamatani, J.; Igawa, S.;
Moriyama, T.; Miura, S.; Takiguchi, T.; Okada, S.; Hoshino, M.; Ueno, K. J. Am. Chem. Soc.,
24
ACS Paragon Plus Environment

Page 25 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
2003, 125, 12971. (d) Chou, P.ꢀT.; Chi, Y. Eur. J. Inorg. Chem. 2006, 3319. (e) Yook, K. S.; Lee, J.
Y. Adv. Mater. 2012, 24, 3169. (f) Sasabe, H.; Kido, J. J. Mater. Chem. C, 2013, 1, 1699.
(
4) (a) Zhang, Q.; Zhou, Q.; Cheng, Y.; Wang, L.; Ma, D.; Jing, X.; Wang, F. Adv. Funct. Mater.
006, 16, 1203. (b) Deaton, J. C.; Switatski, S. C.; Kondakov, D. Y.; Young, R. H.; Pawlik, T. D.;
Giesen, D. J.; Harkins, S. B.; Miller, A. J. M.; Mickenberg, S. F.; Peters, J. C. J. Am. Chem. Soc.
010, 132, 9499. (c) Hashimoto, M.; Igawa, S.; Yashima, M.; Kawata, I.; Hoshino, M.; Osawa, M.
2
2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
J. Am. Chem. Soc. 2011, 133, 10348. (d) Hsu, C.ꢀW.; Lin, C.ꢀC.; Chung, M.ꢀW.; Chi, Y.; Lee,
G.ꢀH.; Chou, P.ꢀT.; Chang, C.ꢀH.; Chen, P.ꢀY. J. Am. Chem. Soc. 2011, 133, 12085. (e) Zhang, Q.;
Komino, T.; Huang, S.; Matsunami, S.; Goushi, K.; Adachi C. Adv. Funct. Mater. 2012, 22, 2327.
(
f) Liu, Z.; Qiu, J.; Wei, F.; Wang, J.; Liu, X.; Helander, M. G.; Rodney, S.; Wang, Z.; Bian, Z.; Lu,
Z.; Thompson, M. E.; Huang, C. Chem. Mater. 2014, 26, 2368.
(5) (a) Blasse, G.; McMillin, D. R. Chem. Phys. Lett. 1980, 70, 1. (b) Kirchhoff, J. R.; Gamache,
R. E.; Blaskie, M. W.; Del Paggio, A. A.; Lengel, R. L.; McMillin, D. R. Inorg. Chem. 1983, 22,
2380. (c) Tsuboyama, A.; Kuge, K.; Furugori, M.; Okada, S.; Hoshino, M.; Ueno, K. Inorg. Chem.
2007, 46, 1992. (d) Czerwieniec, R.; Yu, J.; Yersin H. Inorg. Chem. 2011, 50, 8293.
(6) (a) Goushi, K.; Yoshida, K.; Sato, K.; Adachi, C. Nat. Photonics 2012, 6, 253. (b) Goushi, K.;
Adachi, C. Appl. Phys. Lett. 2012, 101, 023306. (c) Graves, D.; Jankus, V.; Dias, F. B.; Monkman,
A. Adv. Funct. Mater. 2014, 24, 2343. (d) Hung, W.ꢀY.; Fang, G.ꢀC.; Lin, S.ꢀW.; Cheng, S.ꢀH.;
Wong, K.ꢀT.; Kuo, T.ꢀY.; Chou P.ꢀT. Sci. Rep. 2014, 4, 5161.
(7) (a) Nakagawa, T.; Ku, S.ꢀY.; Wong, K.ꢀT.; Adachi C. Chem. Comm. 2012, 48, 9580. (b)
Mehes, G.; Nomura, H.; Zhang, Q.; Nakagawa, T.; Adachi, C. Angew. Chem. Int. Ed. 2012, 51,
11311. (c) Nasu, K.; Nakagawa, T.; Nomura, H.; Lin, C.ꢀJ.; Cheng, C.ꢀH.; Tseng, M.ꢀR.; Yasuda,
T.; Adachi, C. Chem. Commun., 2013, 49, 10385.
(8) (a) Tanaka, H., Shizu, K., Miyazaki, H., Adachi, C. Chem. Comm. 2012, 48, 11392. (b)
Tanaka, H.; Shizu, K.; Nakanotani, H.; Adachi, C. Chem. Mater. 2013, 25, 3766. (c) Lee, S. Y.;
Yasuda, T.; Yang, Y. S.; Zhang, Q.; Adachi, C. Angew. Chem. Int. Ed. 2014, 53, 6402.
(
9) (a) Komino, T.; Tanaka, H.; Adachi, C. Chem. Mater. 2014, 26, 3665. (b) Mayr, C.; Lee, S.
Y.; Schmidt, T. D.; Yasuda, T.; Adachi, C.; Brütting W. Adv. Funct. Mater. 2014, 24, 5232. (c) Sun,
J. W.; Lee, J.ꢀH.; Moon, C.ꢀK.; Kim, K.ꢀH.; Shin, H.; Kim, J.ꢀJ. Adv. Mater. 2014, 26, 5684. (d)
Nakanotani, H.; Higuchi, T.; Furukawa, T.; Masui, K.; Morimoto, K.; Numata, M.; Tanaka, H.;
Sagara, Y.; Yasuda, T.; Adachi, C. Nat. Commun. 2014, 5, 4016. (e) Zhang, D.; Duan, L.; Li, C.; Li,
Y.; Li, H.; Zhang, D.; Qiu, Y. Adv. Mater. 2014, 26, 5050. (f) Nishide, J.ꢀi.; Nakanotani, H.; Hiraga,
Y.; Adachi, C. Appl. Phys. Lett. 2014, 104, 233304. (g) Ishimatsu, R.; Matsunami, S.; Kasahara, T.;
Mizuno, J.; Edura, T.; Adachi, C.; Nakano, K.; Imato, T. Angew. Chem. Int. Ed. 2014, 53, 6993.
(
10) (a) Baldo, M. A.; Adachi, C.; Forrest, S. R., Phys. Rev. B 2000, 62, 10967. (b) Song, D.;
Zhao, S.; Luo, Y.; Aziz, H. Appl. Phys. Lett. 2010, 97, 243304. (c) Masui, K.; Nakanotani, H.;
Adachi, C. Org. Electron. 2013, 14, 2721.
(
11) Turro, N. J.; Ramamurthy, V.; Scaiano, J. C. Modern Molecular Photochemistry of Organic
Molecules; University Science Books, Sausalito, 2010.
(12) (a) Grabowski, Z. R.; Rotkiewicz, K. Chem. Rev. 2003, 103, 3899. (b) Kapturkiewicz, A.;
Herbich, J.; Nowacki J. Chem. Phys. Lett. 1997, 275, 355.
(
13) (a) Newton, M. D. Chem. Rev. 1991, 91, 767. (b) Barbara, P. F.; Meyer, T. J.; Ratner, M. A.
J. Phys. Chem. 1996, 100, 13148.
25
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 26 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(14) (a) Miller, M. T.; Gantzel, P. K.; Karpishin, T. B. J. Am. Chem. Soc., 1999, 121, 4292. (b)
Scaltrito, D. V.; Thompson, D. W.; O’Callaghan, J. A.; Meyer, G. J. Coord. Chem. Rev., 2000,
208, 243. (c) Felder, D.; Nierengarten, J.ꢀF.; Barigelletti, F.; Ventura, B.; Armaroli, N. J. Am.
Chem. Soc. 2001, 123, 6291. (d) Yam, V. W.ꢀW.; Cheng, E. C.ꢀC.; Zhu, N. Chem. Commun., 2001,
37, 1028. (e) Riesgo, E. C.; Hu, Y.ꢀZ.; Bouvier, F.; Thummel, R. P.; Scaltrito, D. V.; Meyer, G. J.
Inorg. Chem. 2001, 40, 3413. (f) Zhang, Q.; Cheng, Y.; Wang, L.; Xie, Z.; Jing, X.; Wang, F. Adv.
Funct. Mater. 2007, 17, 2983. (g) Zink, D. M.; Bächle, M.; Baumann, T.; Nieger, M.; Kühn, M.;
Wang, C.; Klopper, W.; Monkowius, U.; Hofbeck, T.; Yersin, H.; Bräse, S. Inorg. Chem. 2013, 52,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
292. (h) Krylova, V. A.; Djurovich, P. I.; Conley, B. L.; Haiges, R.; Whited, M. T.; Williams, T.
J.; Thompson, M. E. Chem. Commun., 2014, 50, 7176.
15) (a) Plate, M.; Mobius, K.; MichelꢀBeyerle, M. E.; Bixon, M.; Jortner, J. J. Am. Chem. SOC.
1988, 110, 7279. (b) Troisi, A.; Orlandi, G. J. Phys. Chem. B 2002, 106, 2093.
(
(
(
16) ltoh, T. Chem. Rev. 1995, 95, 2351.
17) (a) Runge, E. E.; Gross, K. U. Phys. Rev. Lett. 1984, 52, 997. (b) Marques, M. A. L.; Gross,
E. K. U. Annu. Rev. Phys. Chem. 2004, 55, 427.
(18) (a) Magyar, R. J.; Tretiak, S. J. Chem. Theory Comput. 2007, 3, 976. (b) Harbach, P. H. P.;
Dreuw A. The Art of Choosing the Right Quantum Chemical ExcitedꢀState Method for Large
Molecular Systems. In Modeling of Molecular Properties; Comba, P., Ed.; WileyꢀVCH, 2011; pp
37–38. (c) Guido, C. A.; Knecht, S.; Kongsted, J.; Mennucci B. J. Chem. Theory Comput. 2013, 9,
2209.
(
19) (a) Dreuw, A.; HeadꢀGordon, M. Chem. Rev. 2005, 105, 4009. (b) Peach, M. J. G.; Benfield,
P.; Helgaker, T.; Tozer, D. J. J. Chem. Phys. 2008, 128, 044118. (c) Jacquemin, D.; Perpète, E. A.;
Ciofini, I.; Adamo, C.; Valero, R.; Zhao, Y.; Truhlar, D. G. J. Chem. Theory Comput. 2010, 6,
2071. (d) Sini, G.; Sears, J. S.; Brédas J.ꢀL. J. Chem. Theory Comput. 2011, 7, 602.
(
20) Huang, S.; Zhang, Q.; Shiota, Y.; Nakagawa, T.; Kuwabara, K.; Yoshizawa, K.; Adachi, C.
J. Chem. Theory Comput. 2013, 9, 3872.
21) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Stephens, P. J.; Devlin, F. J.;
Chabalowski, C. F.; Frisch, M. J. J. Phys. Chem. 1994, 98, 11623.
(
(
(
(
22) Lu, T.; Chen, F. W. J. Comp. Chem. 2012, 33, 580.
23) Zhao, Y.; Truhlar, D. G. J. Phys. Chem. A 2004, 108, 6908
24) Boese, A. D.; Martin, J. M. L. J. Chem. Phys. 2004, 121, 3405
(25) Grabowski, Z. R.; Rotkiewicz, K.; Siemiarczuk, A.; Cowley, D. J.; Baumann, W. Nouv. J.
Chim. 1979, 3, 443.
(
26) (a) AlꢀHassan, K. A.; Saleh, N.; AbuꢀAbdoun, I. I.; Yousef, Y. A. J Incl Phenom Macrocycl
Chem. 2008, 61, 361. (b) Koenig, M.; Bottari, G.; Brancato, G.; Barone, V.; Guldi, D. M.; Torres,
T. Chem. Sci. 2013, 4, 2502.
(
27) (a) Maus, M. Photoinduced Intramolecular Charge Transfer in Donor–Acceptor Biaryls
and Resulting Applicational Aspects Regarding Fluorescent Probes and Solar Energy Conversion,
Universal Publishers; USA; 1998. (b) Maus, M.; Rettig, W. Chem. Phys. 1997, 218,151. (c) Maus,
M.; Rettig, W.; Jonusauskas, G.; Lapouyade, R.; Rulliere, C. J. Phys. Chem. A 1998, 102, 7393. (d)
Maus, M.; Rettig, W.; Bonafoux, D.; Lapouyade, R. J. Phys. Chem. A 1999, 103, 3388. (e) Maus,
M.; Rettig, W.; Chem. Phys. 2000, 261, 323.
(28) (a) Davis, W. B.; Ratner, M. A.; Wasielewski, M. R. J. Am. Chem. Soc. 2001, 123, 7877. (b)
Benniston, A. C.; Harriman, A. Chem. Soc. Rev. 2006, 35, 169.
26
ACS Paragon Plus Environment

Page 27 of 28
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
(29) (a) Löwdin, P.ꢀO. Rev. Mod. Phys. 1962, 34, 80. (b) G. Likhtenshtein, Solar Energy
Conversion: Chemical Aspects; WileyꢀVCH, 2012; p 25.
(
30) (a) Improta, R.; Barone, V.; Scalmani, G.; Frisch, M. J. J. Chem. Phys. 2006, 125, 054103.
(
b) Improta, R.; Scalmani, G.; Frisch, M. J.; Barone, V. J. Chem. Phys. 2007, 127,074504.
(31) (a) AbdelꢀShafi, A. A.; Worrall, D. R. J. Photochem. Photobiol. A: Chem. 2005, 172, 170.
b) Li, B.; Nomura, H.; Miyazaki, H.; Zhang, Q.; Yoshida, K.; Suzuma, Y.; Orita, A.; Otera, J.;
(
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Adachi, C. Chem. Lett. 2014, 43, 319.
(32) Klessinger, M.; Michl, J. Excited States and Photochemistry of Organic Molecules; VCH,
Weinheim, 1995; p 247–260.
(
33) (a) Reineke, S.; Seidler, N.; Yost,S. R.; Prins, F.; Tisdale, W. A.; Baldo, M. A. Appl. Phys.
Lett. 2013, 103, 093302. (b) Reineke, S.; Baldo, M. A. Sci. Rep. 2013, 4, 3797.
(34) (a) Togashi, K.; Nomura, S.; Yokoyama, N.; Yasudaabc, T.; Adachi, C. J. Mater. Chem.
2012, 22, 20689. (b) Nakanotani, H., Masui, K., Nishide, J., Shibata, T., Adachi, C. Sci. Rep. 2013,
3
, 2127.
27
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 28 of 28
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table of Contents Graphic
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
28
ACS Paragon Plus Environment