Synthesis and Exciton Dynamics of Donor-Orthogonal Acceptor Conjugated Polymers: Reducing the Singlet-Triplet Energy Gap

Article abstract of DOI:10.1021/jacs.7b03327

The presence of energetically low-lying triplet states is a hallmark of organic semiconductors. Even though they present a wealth of interesting photophysical properties, these optically dark states significantly limit optoelectronic device performance. Recent advances in emissive charge-transfer molecules have pioneered routes to reduce the energy gap between triplets and "bright" singlets, allowing thermal population exchange between them and eliminating a significant loss channel in devices. In conjugated polymers, this gap has proved resistant to modification. Here, we introduce a general approach to reduce the singlet-triplet energy gap in fully conjugated polymers, using a donor-orthogonal acceptor motif to spatially separate electron and hole wave functions. This new generation of conjugated polymers allows for a greatly reduced exchange energy, enhancing triplet formation and enabling thermally activated delayed fluorescence. We find that the mechanisms of both processes are driven by excited-state mixing between π-πand charge-transfer states, affording new insight into reverse intersystem crossing.

Full text of DOI:10.1021/jacs.7b03327

Subscriber access provided by UNIV OF ARIZONA
Article
Synthesis and Exciton Dynamics of Donor-Orthogonal Acceptor
Conjugated Polymers: Reducing the Singlet-Triplet Energy Gap
David M. E. Freeman, Andrew J Musser, Jarvist M. Frost, Hannah L Stern,
Alexander K. Forster, Kealan J. Fallon, Alexandros G Rapidis, Franco Cacialli,
Iain McCulloch, Tracey M. Clarke, Richard H. Friend, and Hugo Bronstein
J. Am. Chem. Soc., Just Accepted Manuscript • Publication Date (Web): 09 Jun 2017
Downloaded from http://pubs.acs.org on June 9, 2017
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 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Synthesis and Exciton Dynamics of Donor-Orthogonal Acceptor
Conjugated Polymers: Reducing the Singlet-Triplet Energy Gap
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
‡
‡
David M. E. Freeman1 , Andrew J. Musser2 †*, Jarvist M. Frost3, Hannah L. Stern2, Alexander K.
Forster1, Kealan J. Fallon1, Alexandros G. Rapidis4, Franco Cacialli4, Iain McCulloch5, Tracey M.
Clarke1, Richard H. Friend2*, Hugo Bronstein1*
1 Department of Chemistry, University College London, 20 Gordon St., London, WC1H 0AJ, UK
2 Cavendish Laboratory, University of Cambridge, JJ Thomson Avenue, Cambridge, CB3 0HE, UK
3 Department of Materials, Imperial College London, Exhibition Road, London, SW7 2AZ, UK
4.Department of Physics and Astronomy and London Centre for Nanotechnology, University College London, Gower Street,
London WC1E 6BT, UK
5 SPERC, King Abdullah University of Science and Technology, Thuwal 23955-6900, Saudi Arabia
ABSTRACT: The presence of energetically low-lying triplet states is a hallmark of organic semiconductors. While they
present a wealth of interesting photophysical properties, these optically dark states significantly limit optoelectronic device
performance. Recent advances in emissive charge-transfer molecules have pioneered routes to reduce the energy gap be-
tween triplets and ‘bright’ singlets, allowing thermal population exchange between them and eliminating a significant loss
channel in devices. In conjugated polymers this gap has proved resistant to modification. Here, we introduce a general
approach to reduce the singlet-triplet energy gap in fully conjugated polymers, using a donor-orthogonal-acceptor motif to
spatially separate electron and hole wavefunctions. This new generation of conjugated polymers allows for a greatly reduced
exchange energy, enhancing triplet formation and enabling thermally-activated delayed fluorescence. We find that the
mechanisms of both processes are driven by excited-state mixing between π-π*and charge-transfer states, affording new
insight into reverse intersystem crossing.
Introduction
kolaenko et al. reduced ΔEST by linking small molecules ca-
pable of thermally activated delayed fluorescence (TADF)
into a non-conjugated polymer host.9 Luo et al. demon-
strated a similar concept with pendant electron-deficient
moieties on a non-conjugated 3,6-polycarbazole back-
bone.10 More recently, Adachi et al. exploited the small
ΔEST of benzophenone to demonstrate TADF in a keto-pol-
ymer.11
The low-energy triplet excited state in conjugated poly-
mers poses a substantial barrier to next-generation optoe-
lectronic device applications. One advantage of conjugated
polymer semiconductors is their strong absorption and
emission, due to the almost total overlap of π and π* and
orbitals. However, the localised and overlapping wavefunc-
tions result in a large triplet-stabilising exchange energy
ΔEST of ≈ 0.7 eV.1-3 The low energy of the non-emissive tri-
plet state means it forms a trap and loss mechanism in vir-
tually all organic optoelectronic devices (OPV, OLED,
OLET, OLASER).4-6 This could be avoided were it possible
to reduce ΔEST through material design. For application to
OLEDs, triplet excitons could be thermally converted back
into singlets via reverse intersystem crossing (RISC), amply
demonstrated in small-molecule systems with small ΔEST.7
In an organic solar cell architecture, substantially reducing
ΔEST would mean that the stochastic generation of triplets
from recombining charges is no longer a loss pathway.7-8
Herein we present a novel concept in polymer design al-
lowing the construction of fully conjugated polymers with
significantly reduced ΔEST (Figure 1). The principles are ap-
plicable to the entire library of existing polymers and we
anticipate these materials will open a new design space in
conjugated polymer synthesis. Conjugated polymers are
often synthesised with linear donor-acceptor combina-
tions to tune the optical energy gap. By removing the ac-
ceptor from the main chain, and placing it physically or-
thogonal to the donor backbone, we achieve a massive re-
duction in the spatial overlap of the frontier molecular or-
bitals, whilst maintaining conjugation along the polymer
backbone. We therefore retain all the benefits of conju-
gated polymers, such as processability and charge
transport along the polymer backbone, while the resulting
materials also benefit from a greatly reduced ΔEST.
Such energy-level manipulation was assumed impossible
in conjugated polymers, due to a consistently observed
ΔEST of ≈ 0.7 eV. Recent work has shown that a reduction
in ΔEST can be achieved in non-conjugated polymers. Ni-
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 9
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 1. Comparison of conventional donor-acceptor (D-A) conjugated polymer and donor-orthogonal acceptor (D-oA) conju-
gated polymer and its effect on the singlet-triplet energy gap.
We demonstrate this concept by modifying two of the
a
a
c
Poly-
mer
Mn
/
Mw
/
PDIa
HO-
MOb
/eV
Egap
ɛ
most common conjugated polymers into donor-orthogo-
nal acceptor (D-oA) materials. Our spectroscopic and com-
putational investigation reveals the interplay of π-π* and
charge-transfer singlet and triplet states that drives triplet
formation and RISC. We synthesized the series of conju-
gated polymers based on established polyfluorene-con-
taining materials shown in Figure 2 (Supporting Infor-
mation S2-S3). The model system SFH is a simple PFO con-
jugated backbone with alternating 9,9′-spirobifluorene
moieties, similar to a previously reported structure.12 In or-
der to create the novel D-oA polymer SFCN, electron with-
drawing cyano groups were introduced onto the orthogo-
nal spirobifluorene unit, a strategy that has been employed
for achieving TADF in small molecules.13-15 Our other
model conjugated polymer incorporating a more electron-
rich backbone co-monomer was ASFH, similar to the well-
known TFB.16 Again, this structure was modified by the in-
troduction of cyano groups on the orthogonal spirobifluo-
rene, yielding the novel D-oA polymer ASFCN.
kDa
kDa
/ eV
/M-
1cm-1
SFH
16
20
18
23
23
29
31
1.44
1.45
1.72
2.13
-5.77
-5.81
-5.37
-5.42
2.97
2.97
2.89
2.79
16104
18384
11844
7841
SFCN
ASFH
ASFCN
49
Table 1. Physical properties of conjugated polymers (a)
Determined by SEC (PS) using PhCl as eluent. (b) Determined
by PESA. (c) Determined from thin film UV-Vis absorption
onset. (d) Determined in CHCl3 solution.
Thin-film absorption spectra (Fig. 3) reveal slight effects
of our D-oA approach on the nature of the absorbing state.
In SFH and SFCN, the primary absorption band closely
matches the parent polymer PFO, with a distinct enhance-
ment of absorption strength in the D-oA material SFCN
versus SFH. The aryl-amine-containing polymers exhibit
narrower band-gaps, and the spectrum of ASFH closely fol-
lows that of isoelectronic TFB. Interestingly, the D-oA ef-
fects are more pronounced when combined with a ‘push’
backbone in ASFCN: a clear red-shift of ~100 meV and re-
duction in extinction coefficient. These effects imply the
absorption has taken on some charge-transfer (CT) char-
acter, as is commonly seen in linear donor-acceptor conju-
gated polymers.and especially ASFCN where the photolu-
minescence is red-shifted, broad and featureless. SFH and
ASFH possessed ionisation potentials in agreement with
their counterparts PFO and TFB. Introduction of the or-
thogonal acceptor groups resulted in a stabilisation of the
HOMO by ~0.05 eV relative to their respective unsubsti-
tuted polymer.
Figure 2. Structures of synthesized conjugated polymers
ACS Paragon Plus Environment

Page 3 of 9
Journal of the American Chemical Society
evolution of spectral shape on these timescales, but instead
1
2
3
4
5
6
7
8
a slight, continuous red-shift with no evidence of a distinct
1(π-π*) (see Fig. S2). The rise in PL intensity evident from
1-3 ns to 5-10 ns is convoluted with the instrument response
but nonetheless reveals that the primary emissive state is
1
not directly photoexcited, i.e. the emissive CT is not the
state that dominates the absorption spectrum but is in-
stead formed though an internal conversion process. It is
significant that the long-lived state is much more strongly
emissive in this polymer. Fluorescence is readily detected
for 100’s of ns, when triplets are the dominant species (see
below). The PL lifetime enhancement for ASFCN is a sig-
nature of delayed fluorescence, which typically arises from
either triplet-triplet annihilation22 or thermally activated
RISC (akin to that proposed for small molecules7). To dif-
ferentiate between the two mechanisms, we measured the
PL decay dynamics at lower excitation density and per-
formed temperature-dependent steady-state measure-
ments.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Figure 3. UV-vis absorption (shaded) and photolumines-
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
cence (unshaded) of polymer thin films. Fluorene-back-
bone polymers SFH and SFCN exhibit the same band gap in
absorption ~417 nm. The electron-rich backbone of ASFH re-
sults in a red-shift of 80 meV to 429 nm, while incorporation
of the polar acceptor units in ASFCN red-shifts the absorption
by a further 100 meV to 444 nm. The effect of -CN substitution
is markedly more pronounced in the excited state, with con-
tributions from charge-transfer emission in SFCN and espe-
cially ASFCN where the photoluminescence is red-shifted,
broad and featureless.
Figure 4. Photoluminescence dynamics in ΔEST-tuned
polymer films. (a) Time-resolved emission spectra of SFCN
film in vacuum, following excitation at 400 nm, with a tem-
poral resolution of ~5 ns. Emission is dominated by singlet (π-
π*) emission. Vertical axis is broken to highlight very weak
contribution from CT singlet emission at 500 nm evident on
longer timescales. Similar behaviour is observed in the ‘push-
only’ polymer ASFH (Fig. S2). (b) Equivalent measurements of
ASFCN films. We detect only broad, structureless CT emis-
sion, with no equivalent to the initial vibronically structured
π-π* emission observed in other polymers. The rise in PL in-
tensity between the first and second spectra in ASFCN reveals
delayed formation of the emissive CT state, though it is con-
voluted with the instrument response. Pump scatter in both
spectral plots is indicated by *. (c) PL decay kinetics measured
for all polymer films, taken at 450 nm (SFH and SFCN) and
500 nm (ASFH and ASFCN). Incorporation of cyanobiflourene
units and introduction of CT states increases the PL lifetime
(light green vs. black, red vs. dark green). In ASFCN, signifi-
cant PL is detected on the 100’s of ns timescale, when triplets
are the dominant species. (d) Steady-state PL spectra of
ASFCN, from 330 K (red) to 77 K (blue). The PL intensity grad-
ually decreases upon cooling, consistent with a major contri-
bution to the PL from TADF. Normalised spectra and equiva-
lent measurement of ASFH in Fig. S3.
Emissive Properties
The effects of D-oA substitution are particularly promi-
nent on the polymer photoluminescence (Fig. 3). SFH and
1
ASFH exhibit predominantly conventional (π-π*) emis-
sion, with a small Stokes shift and well-defined vibronic
progression. Incorporation of cyano groups induces a
slight red-shift and broadening in SFCN, consistent with a
1
new lower energy CT state.17-21 This effect is amplified in
ASFCN, in which the substantial red-shift and absence of
vibronic features indicate that the emission is CT in nature.
In Fig. 4 we illustrate how these spectra evolve in time,
measured on an instrument with a ~5 ns response. ‘Pull-
only’ SFCN exhibits two distinct emissive species, similar
to ‘push-only’ ASFH (Fig. S2): initial vibronically struc-
1
tured (π-π*) emission decays on an instrument-limited
1
timescale to yield a weak, red-shifted defect or CT state
(peak at 500 nm). The latter can be relatively long-lived,
with trace emission detectable up to ~100 ns (Fig. 4c).
Given that the CT state is rapidly populated in SFCN (see
below) yet exhibits very little emission, we conclude that
the 1CT state mixes poorly with the 1(π-π*) and so has neg-
ligible transition dipole moment. ASFCN exhibits no such
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 9
The kinetics shown in Figure 4c were obtained on a 1 kHz
polymers.24 The very large signal suggests a high intersys-
1
2
3
4
5
6
7
8
laser system using amplified pulses. Using an unamplified
2.5 MHz diode laser with dramatically lower pulse ener-
gies, we obtain the same delayed PL decay kinetics in
ASFCN (Fig. S4), demonstrating that the long lifetime is
insensitive to excitation density and should not arise from
triplet-triplet annihilation. As shown in Figure 4d, temper-
ature reduction both decreases the ASFCN thin-film PL
yield and blue-shifts the spectral intensity towards the
peak at 508 nm (Fig. S3). These effects are consistent with
a changing balance of emissive species, with the lower-en-
ergy emitter populated via thermal activation. By contrast,
ASFH exhibits negligible shifts and an increase in PL yield
upon cooling, a typical consequence of suppressing of non-
radiative decay channels (Fig. S3). Our observation of the
opposite behaviour in ASFCN is direct proof of TADF in a
conjugated polymer, enabled by reducing ΔEST through our
D-oA motif. These observations also enable us to rational-
ise the continuous time-resolved PL red-shift in ASFCN.
The thermally activated process will be most efficient at
sites with the lowest ΔEST and thus lowest-energy emissive
CT states, resulting in selective enhancement of the PL life-
time at the red spectral edge. It is interesting to note that,
in contrast to standard small-molecule TADF emitters,
ASFCN exhibits no clear delineation between ‘prompt’ and
‘delayed’ emission.6, 16, 19 Moreover, comparison with tran-
sient absorption data below reveals a strong correlation be-
tem crossing yield, a likely consequence of reducing ΔEST
through our D-oA concept, though in SFCN this effect is
not sufficient to enable RISC. The data in Fig5a thus estab-
1
lish an excited-state pathway in SFCN of (π-π*) → CT →
3(π-π*), in which both conversions are driven by the unique
properties of the D-oA structure.
We observe a similar progression of states in ASFCN in
solution (Fig. 5b): fast 1(π-π*) decay (black, τ ~ 300 fs) yields
an emissive CT state (dark yellow) with PIA below 550 nm.
Following CT decay, a long-lived (τ >3 µs) state with PIA
similar to 3(π-π*) in SFCN is evident (red), which again ex-
hibits strong sensitivity to oxygen (blue dashed kinetic).
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
3
This can be assigned to the (π-π*) state of ASFCN. How-
ever, the distinction between the final states is less sharp:
no 3(π-π*) PIA growth is detected during CT decay, and in-
deed some 3(π-π*) features can be detected throughout the
CT lifetime and vice versa. We consider that the smaller
ΔEST here results in some dynamic equilibrium or mixing
3
between (π-π*) and emissive CT states, analogous to sig-
natures of singlet-mixed triplet-pair states observed in cer-
tain small molecule systems.25-28 Nonetheless, on the long-
est timescales eventual conformational relaxation in the
3(π-π*) makes back-transfer to CT thermally inaccessible
and emission is quenched. Similar results are obtained in
the solid state, with only slight shifts in the characteristic
1
PIA bands of each excited state. In SFCN (Fig. 5c), (π-π*)
3
tween substantial (π-π)* character in the ASFCN excited
→ CT conversion is faster than in solution (τ ~ 400 fs), but
a substantial population of singlet (π-π*) states appear un-
able to form CT and decay with complicated non-exponen-
tial kinetics. This branching of 1(π-π*) decay pathways may
reflect the presence of lower-energy 1(π-π*) ‘trap’ sites aris-
ing from interchain interactions. From the much lower CT
state signal in SFCN film versus solution, we expect that
the 400 fs time constant is dominated by the formation of
these low-energy 1(π-π*) states rather than CT states. Pri-
mary CT state decay is markedly faster (τ ~ 90 ps), though
the relatively small 3(π-π*) signal suggests this is not due to
enhanced intersystem crossing. As previously, the PL
measured on long timescales (open circles, magnified for
clarity) is consistent with the (trace) CT state population,
rather than triplets, confirming that the reduction of ΔEST
in SFCN is insufficient to enable TADF. By contrast,
ASFCN films (Fig. 5d) show stronger evidence of CT-3(π-
π*) mixing than in solution, with features of both states ev-
state and the emission on all timescales, suggesting a more
complex TADF mechanism than is typically proposed.
Photoluminescence quantum yields are discussed in the
supplementary information page S30.
CT-mediated Triplet Formation
To understand the photophysical effects of -CN substi-
tution and the delayed fluorescence mechanism, transient
absorption (TA) spectroscopy was performed. TA is sensi-
tive to the absorption of excited states therefore well-
suited to the study of RISC, where the states of primary in-
terest are optically ‘dark’. In Figure 5a we present the TA
evolution of SFCN in solution. Immediately following pho-
toexcitation, we detect a photo-induced absorption (PIA)
band at 730 nm which can be assigned to the 1(π-π*) (black)
from comparison with the unsubstituted SFH (Fig. S5).
Over ~20 ps this state decays to yield a weaker PIA band at
530 nm (dark yellow); a similar state appears in ASFCN
(Fig. 5b) but not in either reference polymer, indicating it
is a direct consequence of the cyano groups. The decay of
this 530 nm feature in the SFCN data closely tracks the red-
shifted PL detected in solution (Fig. 5a, open circles), and
it can thus be assigned to the SFCN CT state.17, 23 We attrib-
ute the fast formation observed here to the strong electron-
withdrawing character of the cyano groups. On the 10’s of
ns timescale, the spectrum further evolves into a sharp PIA
at 760 nm (red). This final state is remarkably long-lived (τ
> 800 µs), with a lifetime that is highly sensitive to oxygen
(blue dashed kinetic; full data in Fig. S6). These character-
1
ident over the full lifetime following prompt (π-π*) decay
(τ ~ 300 fs). Initially the population balance is primarily CT
(PIA ~550 nm), while beyond 1 ns the 3(π-π*) character (PIA
~750 nm) dominates. Importantly, integrated PL kinetics
(open circles) precisely match the triplet (π-π*) PIA, con-
firming the origin of the red-shifted emission as RISC. The
activation of RISC only in the solid state suggests an im-
portant role for conformational restriction: the 3(π-π*)
state is unable to relax energetically to the same degree as
in solution, maintaining the small CT-3(π-π*) energetic
separation and enabling RISC.
3
istics identify the band at 760 nm as the (π-π*) state, in
agreement with previous studies on similar fluorene-based
ACS Paragon Plus Environment

Page 5 of 9
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 5. Triplet formation in D-oA polymers. TA spectral maps of (a,c) SFCN and (b,d) ASFCN in (a,b) chloroform solution
and (c,d) spin-cast film under vacuum. Spectral cuts (lines) highlight characteristic excited-state features. Kinetics below are in-
tegrated over bracketed regions, with boxes denoting dominant state in that temporal regime. (a) SFCN solution, with clean con-
version from 1(π-π*) (black, 500 fs) to CT (dark yellow, 250 ps) to 3(π-π*) (red, 500 ns). Triplet (lifetime >800 µs) is quenched by
oxygen (dashed kinetic). (b) Similar features identified for ASFCN solution CT (dark yellow, 100 ps) and 3(π-π*) (red, 150 ns) states,
with a broader and shorter-lived 1(π-π*) (black, 300 fs). CT and 3(π-π*) are less distinct, with signatures of each evident throughout
the lifetime of the other. Triplet (lifetime 3.7 µs) is quenched by oxygen (dashed kinetic). (c) SFCN thin film exhibits same 1(π-π*)
(black, 300 fs), CT (dark yellow, 100 ps) and 3(π-π*) (red, 5 ns) characteristics. The 1(π-π*) → CT conversion efficiency and triplet
lifetime are sharply reduced. (d) ASFCN thin films show stronger signs of excited-state mixing. Following prompt 1(π-π*) (black,
300 fs) decay, the population shifts from primarily CT (dark yellow, 1 ns) to primarily 3(π-π*) (red, 150 ns). Open circles: integrated
3
PL kinetics from Figure 4, tracking CT in solution (a,b) and SFCN film (c), and (π-π*) in ASFCN film (d). Break in data 1.5-4 ns
reflects transition from sub-ps-TA (100 fs pump, 400 nm) to ns-TA (1-ns pump, 355 nm), details in Supporting Information.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 9
Mechanism of ΔEST Manipulation
chain, due to relatively weak (albeit non-zero) conjugation
across the spiro linkage.29 The effect of the D-oA motif in
the ground-state geometry is to stabilise the orbitals on the
spiro unit below the π* orbitals along the backbone. Thus,
the first unoccupied state (LUMO) of SFCN and ASFCN is
localised on the orthogonal acceptor group, eliminating
spatial overlap between HOMO and LUMO. Crucially,
spiro-conjugation allows for some electronic communica-
tion across the orthogonal moieties.29 This transition has
very low oscillator strength but affords negligible energetic
separation between S and T excited state energies.
1
2
3
4
5
6
7
8
In small molecules, manipulation of ΔEST to enable RISC
is well known through the study of TADF. The crucial de-
sign rule is the reduction of spatial overlap between occu-
pied and unoccupied molecular orbitals. To investigate
how this rule translates to polymeric systems, we have per-
formed hybrid density functional theory (hybrid DFT,
B3LYP functional) calculations on symmetric oligomers of
our spiro units flanked with hydrogen-capped monomers,
followed by linear response time-dependent density-func-
tional-theory (TDDFT) to probe the excited states. The
small system was chosen for computational efficiency and
the ability to develop a single-electron picture of excita-
tions, rather than for quantitative agreement with experi-
ment. The excited states were well represented by mixes of
the ground-state occupied and virtual Kohn-Sham orbitals,
so we begin by discussing these electron localisations.
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
The Jablonski diagram of excited states as calculated by
TD-DFT (Figure 6b) agree with our orbital analysis. TD-
DFT is an approximate theory. In particular it has a known
difficulty in predicting the relative energy of Singlet and
Triplet states (as the exchange and correlation potential
which leads to this splitting is an uncontrolled approxima-
tion).30 The choice of a TDDFT method is necessitated by
the large size of a minimal model.31 Some of the weakness
of TD-DFT can be sidestepped by applying a further
Tamm-Dancoff approximation (TDA). Thus, the calcula-
tions were carried using LR-TD-DFT at B3LYP/6-31g* (Fig-
ure 6) and 6-311g* both with and without TDA (see sup-
porting information S6). All calculations qualitatively gave
similar orbital energetic orderings.
The major light-absorbing transition in π-conjugated
materials is between delocalised π orbitals and π* excited
states. In typical systems such as SFH/ASFH, these orbitals
are located primarily along the conjugated backbone, with
significant spatial overlap and consequently large oscilla-
tor strength in the singlet configuration (Fig. 6a). In our D-
oA polymers SFCN and ASFCN, we find virtually identical
extension of the π orbital (HOMO) along the polymer
Figure 6. Electronic structure of model oligomers. (a) HOMO (bottom) and LUMO (top) of model trimers calculated using
B3LYP/6-311g*. The D-oA motif results in complete localisation of the LUMO on the orthogonal fluorene unit. (b) Singlet and
triplet states calculated by linear-response time-dependent density-functional-theory (LR-TD-DFT) at B3LYP/6-31g*. The width
of the state indicates the absorption oscillator strength, plotted logarithmically, indicating the large Pi-Pi* transition which will
dominate optical absorption. The numbering is by order of energy, separately for both singlet and triplet states.
ACS Paragon Plus Environment

Page 7 of 9
Journal of the American Chemical Society
The orthogonal acceptor introduces additional charge-
rates and efficiencies of these processes are strongly af-
fected by interchain interactions and restriction of the pol-
ymer conformation and vibrational degrees of freedom in
the solid state,33 although the same basic pathway applies.
Therefore SFCN is not capable of efficient RISC but instead
leads to a novel route to generate polymer triplet excitons.
In ASFCN, we observe signs that triplet formation and, in
thin films, RISC, are mediated by excited-state mixing. The
absorption red-shift and reduction in oscillator strength
upon -CN substitution point to significant CT character in
1
2
3
4
5
6
7
8
transfer states below the main π-π* backbone transition.
This backbone transition remains the strongest absorbing
transition (by ~1000x). Once absorbed there is an energetic
driving force for the backbone π* exciton to transfer to the
orthogonal acceptor, forming a lower energy CT state.
These states are quasi-degenerate triplets and singlets. The
orthogonal alignment of the moieties across the Spiro unit
makes the frontier molecular orbitals (composed of Pi-
character) orthogonal. This lack of molecular orbital over-
lap reduces the oscillator strength of this excited state, but
also reduces the exchange energy that separates states of a
Singlet and Triplet character. Though with TDDFT we can-
not accurately predict the splitting in energy of the Singlet
and Triplet states, the lack of splitting indicates that the
overlap between the hole and electron within the excited
state has tended towards zero, agreeing with our orbital
analysis. In SFCN, these are states S1 and T2, slightly above
the 3(π-π*) of the backbone (T1). In ASFCN, these are states
S1 and T1; the lowest-lying excited states of the material. In
both cases, there is clearly a much richer set of excited
states in the D-oA polymers to facilitate orbital mixing.
Therefore, the weakly donating SFCN polymer retains the
3(π-π*) as the lowest energy excited state meaning that the
singlet-triplet energy gap is reduced to the energetic differ-
9
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
the initial (π-π*) state, as in standard main-chain donor-
acceptor polymers. This state rapidly relaxes into the
lower-energy CT state detected in photoluminescence.
The observed correlation between CT and 3(π-π*) transient
absorption features suggests that this small energetic sep-
aration enables strong mixing between the two states. The
character of this CT-3(π-π*) mixture also evolves slightly
over its lifetime, which we speculate could be attributed to
relaxation of the low-energy vibrational modes coupling
the two potential energy surfaces.34 At no point do we iden-
tify a pure, diabatic triplet or CT configuration, but these
may be transiently populated, indicating the conventional
picture of RISC may not be as simple as suggested.33, 35-36 It
is this mixing with intermediate CT states which enables
RISC to occur: the CT component of CT-3(π-π*) enables a
relatively fast (thermally activated) spin flip, resulting in a
1CT configuration which in turn can emit via mixing with
the 1(π-π*) state. Our results are broadly in agreement with
recent work on small molecule TADF systems which sug-
gest that having a substantial CT contribution to T1 and
that the energetic proximity of a (π-π*) T2 state enhances
RISC.32 Given the success of donor-acceptor motifs to en-
gineer efficient RISC in small molecules, we expect that the
complex mechanism described here may be broadly appli-
cable to other materials capable of TADF
1
ence between the newly created CT state and the previ-
ously existing 3(π-π*). In the case of the stronger donating
ASFCN the newly created singlet and triplet CT states are
3
considerably below the (π-π*), and due to the highly or-
thogonal nature of the spirocyclic charge transfer states we
can assume ΔEST to be very small (<0.1 eV) as has been ob-
served for small molecule spirocyclic TADF materials.32
.
Discussion
Our combined spectroscopic and computational results
demonstrate the utility of our approach to reducing ΔEST
in conjugated polymers. The presence of the spirobifluo-
rene unit on its own does not perturb the photophysics of
the conjugated polymer in any significant manner, as the
HOMO and LUMO are both delocalised along the back-
bone. Incorporation of the strongly electron-withdrawing
-CN groups creates new low-energy CT states which mix to
a varying degree (dependent on ‘push’ from the main
Conclusions
We have demonstrated a concept which allows for the
modification of standard conjugated polymers into ones
with entirely new photophysics. This has been done by in-
troducing orthogonal acceptor groups at the bridgehead
position of alternating fluorene repeat units. This results in
the creation of CT states below the main optically allowed
π-π* transition, dramatically altering the excited state
properties and reducing ΔEST. In SFCN, these new CT
1
chain) with the main-chain (π-π*) state. These CT states
are composed of a delocalised ground state electron on the
polymer backbone and an excited electron localised on the
orthogonal acceptor. Consequently, the exchange energy is
significantly reduced and we observe enhanced triplet for-
mation. In the strongest push-pull D-oA material ASFCN,
these effects combine to enable thermally activated RISC
3
states are higher in energy than the (π-π*) and result in
efficient triplet generation with ultra-long lifetimes, even
at room temperature. Such materials could present an ex-
cellent platform to study the electronic properties of triplet
excitons in more detail, or to harvest them for photochem-
istry such as singlet oxygen generation or photocatalysis.
In the case of ASFCN, these new CT states are below the
3(π-π*) and allow for TADF to occur via RISC, enabling a
detailed transient absorption study of this little-under-
stood phenomenon. Fused aromatic moieties with a
bridgehead, as we have used here, are amongst the most
common synthetic motif in conjugated polymers (eg. fluo-
rene, cyclopentadithiophene, indacenodithiophene). Our
D-oA motif can thus be easily applied to an entire library
3
from the (π-π*) state back into the weakly emissive CT
state when triplet relaxation can be suppressed.
Our results also allow unprecedented insight into the
underlying mechanism of triplet formation and RISC. In
SFCN, in which the energy levels are relatively well sepa-
rated, 1(π-π*) convert predominantly into CT states, allow-
ing for efficient formation of a lower-energy 3(π-π*) states.
1
Residual (π-π*) states decay radiatively, breaking Kasha’s
rule as they are not the lowest excited singlet state. The
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 9
(8)
Wang, J.; Chepelianskii, A.; Gao, F.; Greenham, N. C.,
of structures, leading to a new class of materials with tune-
able ΔEST. A preliminary screen of materials using TD-DFT
shows that using this approach it is possible to fully ma-
nipulate ΔEST across a wide energetic range and to vary the
band-gap across the visible spectrum (Fig. S7). There are
two key advantages to the donor-orthogonal-acceptor de-
sign; firstly it allows for an additional dimension of tuna-
bility (energetically and morphologically) via conjugated
polymer backbone manipulation and secondly, charge
transport routes along the polymer backbone are main-
tained allowing chemists to enter a previously unexplored
design space in conjugated polymer synthesis.
1
2
3
4
5
6
7
8
Nat. Commun. 2012, 3, 1191.
(9)
Roberts, M., Adv. Mater. 2015, 27, 7236.
(10)
mun. 2016, 52, 2292.
(11)
Adv. Mater. 2016, 28, 4019.
(12)
2000, 12, 828.
(13)
1976, 80, 969.
(14)
Org. Electron. 2014, 15, 2027.
(15)
Adachi, C., Angew. Chem. Int. Ed. 2012, 51, 11311.
(16)
krishnan, S.; Zhou, M.; Khong, S.-H.; Chan, H. S. O.; Burroughes,
J. H.; Chua, L.-L.; Friend, R. H.; Ho, P. K. H., Nat. Mater. 2010, 9,
152.
(17)
man, A. P., J. Phys. Chem. C 2007, 111, 18759.
(18) Aydemir, M.; Jankus, V.; Dias, F. B.; Monkman, A. P., J.
Phys. Chem. C 2015, 119, 5855.
Nikolaenko, A. E.; Cass, M.; Bourcet, F.; Mohamad, D.;
Luo, J.; Xie, G.; Gong, S.; Chen, T.; Yang, C., Chem. Com-
Lee, S. Y.; Yasuda, T.; Komiyama, H.; Lee, J.; Adachi, C.,
Yu, W. L.; Pei, J.; Huang, W.; Heeger, A. J., Adv. Mater.
9
Morris, J. V.; Mahaney, M. A.; Huber, J. R., J. Phys. Chem.
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
Méhes, G.; Goushi, K.; Potscavage Jr, W. J.; Adachi, C.,
Méhes, G.; Nomura, H.; Zhang, Q.; Nakagawa, T.;
ASSOCIATED CONTENT
Synthesis and characterization of polymers, additional spec-
troscopic measurements and theoretical calculations. This
material is available free of charge via the Internet at
http://pubs.acs.org.
Png, R.-Q.; Chia, P.-J.; Tang, J.-C.; Liu, B.; Sivarama-
King, S. M.; Hintschich, S. I.; Dai, D.; Rothe, C.; Monk-
AUTHOR INFORMATION
Corresponding Authors
(19)
Hintschich, S. I.; Rothe, C.; King, S. M.; Clark, S. J.;
* a.musser@sheffield.ac.uk
*rhf10@cam.ac.uk
*h.bronstein@ucl.ac.uk
Monkman, A. P., J. Phys. Chem. B 2008, 112, 16300.
(20) dos Santos, P. L.; Dias, F. B.; Monkman, A. P., J. Phys.
Chem. C 2016, 120, 18259.
(21)
Wang, X.; Zhao, L.; Shao, S.; Ding, J.; Wang, L.; Jing, X.;
Present Addresses
Wang, F., Polym. Chem. 2014, 5, 6444.
(22) Wan, Y.; Guo, Z.; Zhu, T.; Yan, S.; Johnson, J.; Huang, L.,
Nat. Chem. 2015, 7, 785.
(23) Devižis, A.; Serbenta, A.; Meerholz, K.; Hertel, D.;
Gulbinas, V., J. Chem. Phys. 2009, 131, 104902.
(24) Rothe, C.; King, S. M.; Monkman, A. P., Phys. Rev. B
2005, 72, 085220.
(25) Stern, H. L.; Musser, A. J.; Gelinas, S.; Parkinson, P.;
Herz, L. M.; Bruzek, M. J.; Anthony, J.; Friend, R. H.; Walker, B. J.,
Proc. Natl. Acad. Sci. 2015, 112, 7656.
(26) Korovina, N. V.; Das, S.; Nett, Z.; Feng, X.; Joy, J.; Haiges,
R.; Krylov, A. I.; Bradforth, S. E.; Thompson, M. E., J. Am. Chem.
Soc. 2016, 138, 617.
(27) Pensack, R. D.; Ostroumov, E. E.; Tilley, A. J.; Mazza, S.;
Grieco, C.; Thorley, K. J.; Asbury, J. B.; Seferos, D. S.; Anthony, J.
E.; Scholes, G. D., J. Phys. Chem. Lett. 2016, 7, 2370.
(28) Bakulin, A. A.; Morgan, S. E.; Kehoe, T. B.; Wilson, M.
W. B.; Chin, A. W.; Zigmantas, D.; Egorova, D.; Rao, A., Nat.
Chem. 2016, 8, 16.
† Dr Andrew Musser
The Department of Physics and Astronomy
Hicks Building
Hounsfield Road
Sheffield, S3 7RH
United Kingdom
Author Contributions
‡These authors contributed equally.
ACKNOWLEDGMENT
Part of this work was funded by EU project 679789 –
CONTREX. AJM was supported by the EPSRC (EP/M01083X).
HLS was supported by the Winton Programme for the Phys-
ics of Sustainability.
(29) Saragi, T. P. I.; Spehr, T.; Siebert, A.; Fuhrmann-Lieker,
T.; Salbeck, J., Chem. Rev. 2007, 107, 1011.
(30) Hait, D.; Zhu, T.; McMahon, D. P.; Van Voorhis, T., J.
Chem. Theory Comput. 2016, 12, 3353.
REFERENCES
(1)
Köhler, A.; Beljonne, D., Adv. Funct. Mater. 2004, 14, 11.
(2)
Monkman, A. P.; Burrows, H. D.; Hartwell, L. J.; Hors-
(31)
Petersilka, M.; Gross, E. K. U.; Burke, K., Int. J. Quantum
burgh, L. E.; Hamblett, I.; Navaratnam, S., Phys. Rev. Lett. 2001,
86, 1358.
Chem 2000, 80, 534.
(32) Samanta, P. K.; Kim, D.; Coropceanu, V.; Brédas, J.-L., J.
Am. Chem. Soc. 2017, 139, 4042.
(33) Gibson, J.; Monkman, A. P.; Penfold, T. J., Chem-
PhysChem 2016, 17, 2956.
(34) Etherington, M. K.; Gibson, J.; Higginbotham, H. F.;
Penfold, T. J.; Monkman, A. P., Nat. Commun. 2016, 7, 13680.
(35) Zhang, Q.; Li, B.; Huang, S.; Nomura, H.; Tanaka, H.;
Adachi, C., Nat. Photon. 2014, 8, 326.
(3)
Andernach, R.; Utzat, H.; Dimitrov, S. D.; McCulloch, I.;
Heeney, M.; Durrant, J. R.; Bronstein, H., J. Am. Chem. Soc. 2015,
137, 10383.
(4)
Rao, A.; Chow, P. C. Y.; Gelinas, S.; Schlenker, C. W.; Li,
C.-Z.; Yip, H.-L.; Jen, A. K. Y.; Ginger, D. S.; Friend, R. H., Nature
2013, 500, 435.
(5)
bley, S.; Thompson, M. E.; Forrest, S. R., Nature 1998, 395, 151.
(6) Kasemann, D.; Brückner, R.; Fröb, H.; Leo, K., Phys. Rev.
B 2011, 84, 115208.
Baldo, M. A.; O'Brien, D. F.; You, Y.; Shoustikov, A.; Si-
(36) Jankus, V.; Aydemir, M.; Dias, F. B.; Monkman, A. P.,
Adv. Sci. 2016, 3, 1500221.
(7)
Uoyama, H.; Goushi, K.; Shizu, K.; Nomura, H.; Adachi,
C., Nature 2012, 492, 234.
ACS Paragon Plus Environment

Page 9 of 9
Journal of the American Chemical Society
Insert Table of Contents artwork here
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
9
ACS Paragon Plus Environment