Solid film versus solution-phase charge-recombination dynamics of exTTF-bridge-C60 dyads

Article abstract of DOI:10.1002/chem.200401312

The charge-recombination dynamics of two exTTF-C₆₀ dyads (exTTF = 9,10-bis 1,3-dithiol-2-ylidene)-9,10-dihydroanthracene), observed after photoinduced charge separation, are compared in solution and in the solid state. The dyads differ only in

Full text of DOI:10.1002/chem.200401312

DOI: 10.1002/chem.200401312
Solid Film versus Solution-Phase Charge-Recombination Dynamics of
exTTF–Bridge–C60 Dyads**
Samantha Handa,[a] Francesco Giacalone,[b] Saif A. Haque,[a] Emilio Palomares,*[a, c]
Nazario Martín,*[b] and James R. Durrant[a]
Abstract: The charge-recombination
dynamics of two exTTF–C60 dyads
(exTTF=9,10-bis(1,3-dithiol-2-yli-
ger electronic coupling in the conjugat-
ed dyad. However, in solid films, the
dynamics are remarkably different,
with dyad 2 showing slower recombina-
tion dynamics than 1. For dyad 1, re-
combination dynamics for the solid
films are observed to be tenfold faster
than in solution, with this acceleration
attributed to enhanced electronic cou-
pling between the geminate radical
pair in the solid film. In contrast, for
dyad 2, the recombination dynamics in
the solid film exhibit a lifetime of 7 ms,
tenfold slower than that observed for
this dyad in solution. These slow re-
combination dynamics are assigned to
the dissociation of the initially formed
geminate radical pair to free carriers.
Subsequent trapping of the free carri-
ers at film defects results in the ob-
served slow recombination dynamics. It
is thus apparent that consideration of
solution-phase recombination data is of
only limited value in predicting the
solid-film behaviour. These results are
discussed with reference to the devel-
opment of organic solar cells based
upon molecular donor–acceptor struc-
tures.
dene)-9,10-dihydroanthracene),
ob-
served after photoinduced charge sepa-
ration, are compared in solution and in
the solid state. The dyads differ only in
the degree of conjugation of the bridge
between the donor (exTTF) and the
acceptor (C60) moieties. In solution,
photoexcitation of the nonconjugated
dyad C60–BN–exTTF (1) (BN=1,1’-bi-
naphthyl) shows slower charge-recom-
bination dynamics compared with the
conjugated dyad C60–TVB–exTTF (2)
(TVB=bisthienylvinylenebenzene)
Keywords: C60 dyads
·
donor–
acceptor systems · electron transfer ·
self-assembly · thin films
(lifetimes of 24 and 0.6 ms, respective-
ly), consistent with the expected stron-
Introduction
The achievement of ultrafast photoinduced electron transfer
from conjugated polymers to C60 moieties has led to novel
approaches for the fabrication of photovoltaic devices. Or-
ganic photovoltaic devices based upon an interpenetrating
network of light-absorbing semiconducting polymer and
electron-accepting C60-derivative materials have resulted in
device efficiencies in excess of approximately 3%.[1] At
present, such interpenetrating networks typically comprise a
random blending of the electron-donor and -acceptor spe-
cies, with blend-morphology optimisation largely achieved
through empirical correlations between film-deposition con-
ditions and device performance. Control of this blend mor-
phology on the nanometer scale is essential to achieve effi-
cient charge separation and charge transport to the device
contacts, whilst at the same time minimising interfacial
charge-recombination losses. There is therefore considerable
interest in developing approaches to control the morphology
of the materials on the nanoscale. One elegant approach to
[a] S. Handa, S. A. Haque, Dr. E. Palomares, J. R. Durrant
Centre for Electronic Material and Devices
Department of Chemistry, Imperial College
Exhibition Road, South Kensington, SW7 2AZ, London (UK)
Fax : (+44)207-594-5801
E-mail: emilio.palomares@uv.es
[b] F. Giacalone, N. Martín
Departamento de Química Orgánica
Facultad de Ciencias Químicas
Universidad Complutense, 28040 Madrid (Spain)
Fax : (+34)91-394-4103
E-mail: nazmar@quim.ucm.es
[c] Dr. E. Palomares
Present address: Institut de Ciencia Molecular (IcMol)
Universitat de Valencia, 46100 Burjassot, Valencia (Spain)
Fax : (+34)96-354-4859
[**] exTTF=9,10-bis(1,3-dithiol-2-ylidene)-9,10-dihydroanthracene, that
is, p-extended tetrathiafulvalene.
7440
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 7440 – 7447

FULL PAPER
this problem that is currently receiving attention is the use
of supermolecular donor–acceptor dyads. In such dyads, the
donor and acceptor species are covalently bound, attached
by a suitable molecular bridge such as to allow charge sepa-
ration whilst minimising the charge-recombination back re-
action.[2–4] Further studies have addressed the potential for
engineering such dyads to allow self-assembly and form or-
dered solid-state materials.[5,6]
Supermolecular and supramolecular donor–acceptor
structures have been widely studied in solution as simple
model systems for studies of photoinduced electron-transfer
dynamics.[7,8] Careful design of molecular structure has led
to near-unity charge-separation yields, with lifetimes of the
charge-separated state of up to hundreds of microseconds,
even for simple dyad structures.[9–11] Moreover, this molecu-
lar approach has been employed for photochemical energy
conversion including, for example, the harnessing of solar ir-
radiation for the generation of adenosine triphosphate
(ATP) in solution.[12] However, the application of such mo-
lecular structures to organic photovoltaics requires consider-
ation and optimisation of their behaviour in solid films. Sev-
eral studies have addressed the photovoltaic performance of
organic thin films fabricated from C60–OPV (fullerene–oli-
go(phenylenevinylene))[13–20] and C60–OPE (fullerene–oligo-
(phenyleneethynylene))[21,22]conjugates and double cables.
However, only very modest device efficiencies were ach-
ieved that were attributed to excessive charge-recombina-
tion losses.[10]
Recently, Janssen et al. have reported a comparison of
electron-transfer dynamics for a C60–OPV dyad in solution
and as a solid film.[23] This study focused on the ultrafast dy-
namics of this system and found that this dyad exhibited
both faster charge-separation and slower charge-recombina-
tion dynamics in the solid state relative to the solution state.
The faster charge-separation dynamics were assigned to in-
termolecular electron-transfer dynamics, whilst the slower
recombination dynamics were assigned to the migration of
charge carriers to energetically more favourable sites within
the film. We believe that the optimisation of photovoltaic
devices based upon molecular donor–acceptor structures re-
quires a detailed understanding of the correlation between
molecular structure and electron-transfer dynamics of these
structures in solid films. For this reason, we begin this report
by making a detailed comparison of the charge-recombina-
tion dynamics of two molecular dyads, C60–BN–exTTF (1)
(BN=1,1’-binaphthyl) and C60–TVB–exTTF (2) (TVB=bis-
thienylvinylenebenzene), in which the donor (exTTF) and
acceptor (C60) units are connected by a chiral binaphthyl p-
conjugated system. These dyads are investigated in both so-
lution and solid films.
Figure 1. Chemical structures of C60–BN–exTTF (dyad 1) and C60–TVB–
exTTF (dyad 2).
cationic species.[24–33] Moreover, careful design of the molec-
ular dyads has enabled us to prepare moieties with long
alkyl side chains attached to the molecular bridge to allow
high solubility in organic solvents for both systems. Further-
more, dyad 1 possesses a binaphthyl unit as molecular
bridge, which breaks the molecular conjugation between
donor and acceptor moieties.[34] In contrast, for dyad 2, the
p-orbital conjugation extends from the exTTF group to the
C60 unit by using a bisthienylvinylenebenzene (o-1,4-dia-
lkoxy-2,5-bis[2-(2-thienyl)vinyl]benzene) unit. We attempt
to rationalise the different dynamics we observe in solution
and solid films in terms of the molecular structures of the
dyads and discuss these results in terms of the optimisation
of dyad structures for photovoltaic device applications.
Experimental Section
Synthesis of molecular dyads
General: FTIR spectra were recorded with KBr pellets on a Nicolet-
Magna-IR 5550 spectrometer. Electrospray ionization (ESI) mass spectra
were recorded on a HP1100MSD spectrometer. UV/Vis spectra were re-
corded in dichloromethane using 1 cm quartz cuvettes on a Varian Cary
50 Scan spectrophotometer. NMR spectra were recorded on Bruker AC-
200 (1H, 200 MHz; 13C, 50 MHz) and AMX-500 (1H, 500 MHz; 13C,
125 MHz) spectrometers at 298 K using partially deuterated solvents as
internal standards. Chemical shifts are given as d values (internal stan-
dard: TMS). Elemental analyses were performed on Perkin–Elmer 2400
CHN and 2400 CHNS/O analysers. Cyclic voltammograms were recorded
on a potentiostat/galvanostat AUTOLAB with PGSTAT30 equipped
with GPES software for Windows (version 4.8). A conventional three-
compartment cell was used with a glassy carbon electrode (GCE) as the
working electrode, a saturated calomel electrode (SCE) as the reference
electrode, Pt wire as the counter electrode, Bu4NClO4 as the supporting
electrolyte, an o-dichlorobenzene/acetonitrile (DCB/MeCN) solvent mix-
ture (4:1 v/v), and a scan rate of 200 mVsꢀ1 (also see Table 1 below).
Figure 1 illustrates the chemical structures of the dyads
used in this study, which differ only in the degree of conju-
gation of the molecular wire that connects both moieties.
The use of tetrathiafulvalene and p-extended tetrathiafulva-
lene (exTTF) units is based on their strong electron-donat-
ing character and their thermodynamic stability when oxi-
dised, which stems from the gain in aromaticity of the di-
Aldehyde 9: A mixture of the p-extended TTF derivative 8[35] (112 mg,
0.15 mmol) and potassium tert-butoxide (34 mg, 0.3 mmol) was placed
under reflux in dry toluene under an argon atmosphere for 30 minutes.
Once the formation of the ylide was completed, a solution of 1,4-dihex-
Chem. Eur. J. 2005, 11, 7440 – 7447
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7441

E. Palomares, N. Martín et al.
yloxy-2,5-bis[(1E)-2’(5-formyl-2-tienyl)vinyl]benzene
(7)[36]
(117 mg,
cleaned by using mixtures of acetonitrile/distilled water/isopropanol and
sonicated for 30 minutes. After the cleaning step, the glass was dried at
1208C for 30 minutes to ensure removal of all solvents. Thin films were
obtained from a diluted solution of each molecular dyad. Films were fab-
ricated for a range of different solvents (dichloromethane, chloroben-
zene, toluene and acetonitrile) and dyad concentrations (1–15×10ꢀ5 m) by
employing both drop-casting and spin-coating methods. Optimum optical
quality of the deposited films was obtained with dichloromethane. The
kinetics of the transient optical data was found to be independent of
dyad concentration, deposition technique and the use of sonication prior
to film deposition. All data reported herein were obtained with films
0.22 mmol) in toluene (10 mL) was added in one portion and the mixture
was reacted further under reflux conditions for 3 h. The crude product
was cooled to room temperature and CH3OH (5 mL) was added. After
evaporation of the solvent mixture, the residue was purified by chroma-
tography on silica gel with a hexane/CH2Cl2 (3:2) mixture as eluent to
give the corresponding dyad as a red solid (63 mg, 45%). 1H NMR
ꢀ
(200 MHz, CDCl3, 258C, TMS): d=9.85 (s, 1H; CHO), 7.81 (s, 1H),
7.73–7.65 (m, 4H), 7.50 (d, 3J(H,H)=16.4 Hz, 2H), 7.34 (d, 3J(H,H)=
16.1 Hz, 1H), 7.33 (d, 3J(H,H)=16.5 Hz, 1H), 7.32 (m, 2H), 7.23 (d, 3J-
(H,H)=16.4 Hz, 2H), 7.14 (m, 2H), 7.04 (s, 2H), 6.94 (m, 3H), 6.32 (s,
drop cast from dichloromethane with dyad concentrations of 1×10ꢀ5
unless otherwise stated.
m
3
ꢀ
ꢀ
4H), 4.05 (t, J(H,H)=6.2 Hz, 4H), 1.89 (m, 4H; CH2 ), 1.55–1.15 (m,
4H), 1.25 (s, 4H), 0.88 ppm (m, 6H; CH3); 13C NMR (50 MHz, CDCl3,
ꢀ
258C, TMS): d=182.47, 153.69, 151.70, 151.03, 142.76, 142.25, 141.26,
137.27, 135.88, 135.80, 135.30, 135.25, 134.82, 134.69, 128.29, 128.18,
127.94, 127.32, 127.23, 125.99, 125.35, 125.06, 124.96, 124.15, 123.33,
123.02, 122.60, 122.09, 121.90, 120.98, 117.27, 117.07, 111.13, 110.38, 69.62,
69.42, 31.65, 31.59, 29.44, 29.37, 25.95, 25.92, 22.68, 22.64, 14.11,
14.04 ppm; IR (KBr) nΡ=2925, 2854, 1655, 1493, 1464, 1433, 1377, 1260,
1219, 1043, 943, 798, 635 cmꢀ1; UV/Vis (CH2Cl2): lmax (loge)=233 (4.60),
367 (4.40), 479 nm (4.80 molꢀ1 cm3 dmꢀ1); MS (ESI): m/z (%): 949 (100)
[M+Na]+.
Optical transient studies: Nanosecond–microsecond transient spectrosco-
py experiments were performed by using a Xenon lamp as a probe
source and a Nd-YAG laser as excitation source (lex =335 nm and pulse
duration <6 ns) at 1 Hz. The resulting photoinduced change in absorp-
tion was monitored by employing a 75 W Xenon arc lamp, a PTI model
101 monochromator (dual grating) after the sample, a TDS Tektronix
2022 digital storage oscilloscope and an Si-based photodiode (Costronics
Electronics) as a photodetector. Data acquisition was carried out by
using Tekave version 1.43 software. The UV/Vis spectra before and after
the laser transient experiments were measured by using a double-beam
Shimadzu UV-1601 spectrophotometer.
Dyad 2: A mixture of the aldehyde 9 (45 mg, 0.049 mmol), [60]fullerene
(35 mg, 0.049 mmol) and N-octylglycine (27 mg, 0.15 mmol) in chloroben-
zene (28 mL) was placed under reflux for 24 hours. After cooling to
room temperature, the crude product was purified by column chromatog-
raphy on silica gel, using CS2 to elute the unreacted fullerene, followed
by a hexane/toluene (7:3) mixture to isolate compound 2 as a black solid
(39 mg, 45%). M.p. 206–2088C (hexane/toluene); 1H NMR (500 MHz,
CDCl3, 258C, TMS): d=7.80 (d, 3J(H,H)=1.4 Hz, 1H), 7.70 (m, 2H),
7.67 (d, J=8.0 Hz, 1H), 7.35 (dd, 3J1(H,H)=8.0 Hz, 3J2(H,H)=1.4 Hz,
1H), 7.30 (d, 3J(H,H)=3.4 Hz, 1H), 7.29 (d, 3J(H,H)=3.4 Hz, 2H), 7.23
(d, 3J(H,H)=16.5 Hz, 2H), 7.23–7.16 (m, 3H), 7.01 (s, 1H), 6.98 (d, 3J-
Results
Synthesis: Dyad 1 (exTTF–BN–C60) was synthesized by fol-
lowing the method previously reported by our group.[34]
Dyad 2 (exTTF–TVP–C60) was obtained in a multistep syn-
thetic procedure as depicted in Scheme 1. Thus, a twofold
Wittig–Horner olefination reaction of bisphosphonate 3[37]
with commercially available 4-bromo-2-formylthiophene
3
3
(H,H)=16.5 Hz, 2H), 6.96 (d, J(H,H)=16.5 Hz, 1H), 6.95 (d, J(H,H)=
16.5 Hz, 1H), 6.94 (m, 1H), 6.31 (s, 4H), 5.33 (s, 1H), 5.08 (d, 3J(H,H)=
9.6 Hz, 1H), 4.10 (d, 3J(H,H)=9.6 Hz, 1H), 4.02 (m, 4H), 3.43 (m, 1H),
2.62 (m, 1H), 1.87 (m, 6H), 1.55–1.25
(m, 22H), 0.93 ppm (t, 9H); 13C NMR
(125 MHz, CDCl3, 258C, TMS): d=
156.19, 154.22, 153.38, 153.34, 151.13,
151.07, 147.32, 147.31, 146.94, 146.37,
146.31, 146.27, 146.19, 146.15, 146.12,
146.07, 145.94, 145.92, 145.78, 145.51,
145.50, 145.46, 145.42, 145.32, 145.28,
145.24, 145.15, 144.85, 144.71, 144.65,
144.38, 144.36, 143.15, 143.02, 142.96,
142.68, 142.56, 142.32, 142.23, 142.14,
142.07, 142.03, 141.96, 141.94, 141.90,
141.65, 141.59, 140.15, 140.12, 139.90,
139.73, 137.08, 136.69, 135.85, 135.80,
135.73, 135.56, 135.28, 135.24, 134.73,
129.02, 128.69, 128.21, 128.07, 127.34,
126.87, 126.43, 126.35, 125.98, 125.33,
125.29, 125.08, 124.96, 124.93, 124.13,
123.70, 123.63, 122.57, 122.43, 122.28,
122.08, 121.97, 117.32, 117.25, 117.10,
110.93, 110.48, 109.57, 78.49, 69.53,
69.45, 68.69, 66.97, 53.60, 53.40, 31.97,
31.65, 29.71, 29.45, 29.41, 29.35, 28.32,
27.56, 25.95, 22.73, 22.71, 22.69, 14.18,
14.13 ppm; IR (KBr): nΡ=2946, 2920,
2851, 1635, 1541, 1451, 1183, 941, 752,
635, 526 cmꢀ1; UV/Vis (CH2Cl2): lmax
(loge)=256 (5.21), 309 (4.76), 332
(4.77), 441 (4.92, sh), 468 nm
(4.96 molꢀ1 cm3 dmꢀ1); MS (ESI): 1773
(100) [M+H]+.
Thin film preparation: All glass used
for drop-casting of solutions was
Scheme 1. Synthesis of the C60–TVB–exTTF dyad (2).
7442
www.chemeurj.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 7440 – 7447

exTTF–Bridge–C60 Dyads
FULL PAPER
under basic conditions afforded the p-conjugated compound
5 bearing two bromine atoms at terminal positions. A fur-
ther Rosemund–von Braun reaction with copper(i) cyanide
in DMF, followed by treatment with ammonia, led to the di-
cyano derivative 6[38]. Compound 6 underwent a reduction
process with DIBAL-H and hydrolysis to form dialdehyde 7.
A careful stoichiometric Wittig reaction of 7 with exTTF
phosphonium salt 8[36] afforded compound 9 bearing the
exTTF unit and a suitable formyl group. This formyl group
reacted with C60 in the presence of N-octyl glycine to form
the final fulleropyrrolidine 2, according to the Prato proce-
dure,[39] by 1,3-dipolar cycloaddition of the azomethyne
ylide, generated in situ, to the C60 molecule (Scheme 1).
Purification of 2 was accomplished by using flash chroma-
tography (silica gel, CS2 to recover unreacted C60, then
hexane/toluene 7:3) to obtain dyad 2 in 45% yield. The
1H NMR spectrum (500 MHz) shows, in addition to the
presence of the protons of exTTF and the pyrrolidine ring,
the presence of the all-trans vinyl groups at d=7.23, 6.98,
6.96 and 6.95 ppm (J3 ꢁ16.5 Hz, see the Experimental Sec-
tion). The UV-visible spectrum of 2 shows that it consists of
a simple superimposition of the constituent fragments, thus
confirming the lack of significant electronic interactions be-
tween the electroactive species in the ground state.
signed to the reduction of the p-conjugated bridge (see
Figure 2). On the oxidation side, important differences are
found between dyads 1 and 2. Thus, dyad 1 shows a first
Figure 2. Cyclic voltammograms for dyad 2 at room temperature (solvent,
oDCB/MeCN 4:1 v/v; supporting electrolyte, Bu4NClO4, scan rate,
200 mVsꢀ1).
quasireversible oxidation wave corresponding to the forma-
tion of the dication of the exTTF unit at +0.57 V, with two
additional waves at +1.04 and +1.33 V stemming from the
BN moiety. Dyad 2 also exhibits the quasireversible oxida-
tion wave of the exTTF unit and two further oxidation
waves corresponding to the p-conjugated bridge (TVB).
However, the first oxidation potential value in 2 is signifi-
cantly shifted towards less positive values relative to the
parent exTTF (Table 1). This cathodic shift is in contrast to
that observed for dyad 1 and reveals the better electronic
communication existing between the exTTF unit and the C60
moiety through the p-conjugated bridge in dyad 2. This find-
ing is in agreement with the behaviour previously observed
for the related C60–TVB dyads, which also showed a small
but noticeable electronic interaction between the p-conju-
gated oligomer and C60.[38]
Electrochemical characterisation: The electrochemical fea-
tures of dyads 1 and 2 were probed by using cyclic voltam-
metry in solution at room temperature, and showed ampho-
teric redox behaviour with donor and acceptor electroactive
responses. The redox potentials are collected in Table 1 with
those of the parent exTTF, pristine C60 and the unsubstitut-
ed fulleropyrrolidine (Fp) for comparison (Table 1).
Table 1. Redox potential (E in V) values of novel compounds 1 and 2
and reference compounds.[a,b]
E1pa
E2pa
E3pa
E1pc
E2pc
E3pc
E4pc
E5pc
exTTF +0.55
–
–
–
–
–
–
–
–
–
–
–
–
–
–
C60
Fp
1
–
–
ꢀ0.54 ꢀ0.96 ꢀ1.43 ꢀ1.92
Behaviour in solution: First we consider the behaviour of
dyads 1 and 2 in solution. Figure 3 shows the absorption
spectra of both dyads in solution (black lines). As expected,
both molecular dyads exhibit a strong absorption maximum
in the blue spectral region, with the absorption maximum of
dyad 2 redshifted relative to that of dyad 1 (labs =465 and
434 nm, respectively), consistent with the expected greater
delocalisation of the p orbitals for this dyad.
ꢀ0.64 ꢀ1.03 ꢀ1.58 ꢀ1.99
+0.57 +1.04 +1.33 ꢀ0.64 ꢀ1.05 ꢀ1.60 ꢀ2.09
2
+0.50 +0.91 +1.29 ꢀ0.66 ꢀ1.07 ꢀ1.62 ꢀ1.87 ꢀ2.05
[a] Conditions: V versus SCE; working electrode, GCE; reference elec-
trode, Ag/Ag+; counter electrode, Pt; 0.1m Bu4NClO4; scan rate,
200 mVsꢀ1; concentrations, 0.5–2.0×10ꢀ3 m; solvent, oDCB:MeCN (4:1
v/v). [b] Epaand Epc are anodic and cathodic peak potentials, respectively.
The primary function of the molecular wire bridge in the
dyads is to allow photoinduced electron transfer from the
exTTF donor unit to the C60 electron-accepting molecule,
thereby resulting in a long-lived charge-separated radical-
pair state. We herein employ transient absorption spectros-
copy to monitor these electron-transfer dynamics by focus-
ing, in particular, upon the lifetime of the charge-separated
state. A long lifetime of this state is essential for photovolta-
ic device function, allowing efficient charge collection by the
device electrodes. Figure 4 compares typical data for both
dyads in solution, with Figure 5 showing the transient ab-
Similarly to dyad 1,[34] dyad 2 shows the presence of four
quasireversible one-electron reduction waves, which corre-
spond to the reduction of the fullerene core at ꢀ0.66, ꢀ1.07,
ꢀ1.62 and ꢀ2.05 V. These values are cathodically shifted rel-
ative to pristine C60 and have been explained in terms of the
saturation of a double bond of the fullerene core, which
raises the LUMO energy. These values are furthermore
quite similar to those found for the unsubstituted fulleropyr-
rolidine (Fp). However, in contrast to dyad 1, another re-
duction wave is observed in dyad 2 at ꢀ1.87 V, which is as-
Chem. Eur. J. 2005, 11, 7440 – 7447
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7443

E. Palomares, N. Martín et al.
Figure 5. Transient absorption spectra (lex =355 nm) recorded after t=
1 ms for dyad 1 in dichloromethane (top) and on thin film (bottom). The
spectra show the features corresponding to exTFFC+ (lꢁ700 nm) and
C60Cꢀ (lꢁ980 nm) for both samples.
sorption spectrum observed one microsecond after pulsed
laser excitation at lex =355 nm.
Figure 3. The UV-visible spectra of dyad 1 (top) and dyad 2 (bottom) in
dichloromethane (black curves) and film (grey curves), spin coated on
transparent glass. The insets show the AFM images of the films. Note the
AFM and the optical spectra both indicate an aggregated, more scattered
film for dyad 2 relative to 1.
Both dyads 1 and 2 show the expected formation of an
exTTF radical cation (exTTFC+)/C60 radical anion (C60Cꢀ)
charge-separated state. In both cases, photoinduced absorp-
tion maxima are observed at lꢁ700 nm, assigned to the
+
exTTFC species, and at lꢁ980 nm, assigned to the C60Cꢀ
species. These values are consistent with previous observa-
tions.[12]
Typical spectra for charge recombination of the exTTFC+/
C60Cꢀ state in solution are shown in Figure 5, monitored at l
ꢁ980 nm (assigned to the C60Cꢀ absorption). It is apparent
that the decay dynamics of 1 are retarded compared with
those of 2, with decay lifetimes, t50%, of t=25 and 0.7 ms, re-
spectively. This retardation is consistent with the expected
less-conjugated character of the molecular bridge of dyad 1.
The faster recombination for dyad 2 results from the extend-
ed conjugation of the molecular bridge “wiring” the donor
and acceptor species, thereby enhancing the electronic cou-
pling between these species.
For both dyads, the recombination dynamics are nonexpo-
nential, as is apparent from their dispersive appearance on
the logarithmic timescale employed. Numerical fitting of the
decay dynamics to a stretched exponential function (DOD/
exp(ꢀ(t/t)a) gave reasonable fits to the data, with stretch pa-
rameter aꢁ0.45 for both dyads. This nonexponential behav-
iour is tentatively assigned to different structural conforma-
Figure 4. Log–linear plot for the electron-recombination dynamics for
dyads 1 (grey) and 2 (black) in dichloromethane at lprobe =830 and lex
=
355 nm, with their representative stretched exponential fits. The inset
shows the linear–linear plot.
7444
www.chemeurj.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 7440 – 7447

exTTF–Bridge–C60 Dyads
FULL PAPER
tions of the dyads, as discussed in more detail below. Fur-
thermore, we would like to point out that despite the use of
diluted solutions (1×10ꢀ5 m) for both dyads, we still ob-
served aggregation processes for dyad 2. This behaviour im-
plies that, although experimental conditions are identical for
both samples, the nature of the molecular bridge influences
the formation of aggregates.
Behaviour in the solid state: We now turn to the investiga-
tion of molecular films fabricated by drop casting dyads 1
and 2 onto glass substrates from solution. For dyad 1, the
ground-state absorption spectra in solution and in the solid
film were indistinguishable, indicating that exTTF p orbitals
involved in the optical absorption were not strongly pertur-
bed in the film relative to their state in solution. The optical
quality of the films formed with dyad 2 was found to be
poorer than that for dyad 1, resulting in significant light
scattering and preventing the determination of a precise ab-
sorption spectrum for this dyad in the solid film. The pro-
nounced light scattering of films fabricated from dyad 2 is
attributed to the formation of molecular aggregates, as dis-
cussed below.
Figure 6 compares transient absorption data obtained for
dyad 1 in solution and as a solid film. For such film studies,
it is essential to address the excitation-density dependence
of the transient data. For isolated molecules in solution, the
transient kinetics are expected to be excitation-density inde-
pendent. However, for solid films, nonlinear effects due to
polaron–polaron, exciton–exciton or exciton–polaron inter-
actions can arise at higher excitation densities, resulting in
excitation-density-dependent dynamics and complicating in-
terpretation of the experimental data. For this reason, tran-
sient absorption data were collected as a function of excita-
tion density. Figure 6 (top) compares dynamics for solution
and solid films at two (high and low) excitation densities,
whilst Figure 6 (bottom) shows the dependence of the am-
plitude of the transient signal on laser intensity for both
samples. Most importantly, it is apparent that the recombi-
nation dynamics for the solid film are approximately tenfold
faster than those observed in solution (lifetimes of t=2.1
and 25 ms). It is furthermore apparent that the charge-re-
combination kinetics for both the solution and film are inde-
pendent of excitation density over the range studied. The
magnitude of the transient signal exhibits a linear depend-
ence on excitation density for the solution data as expected,
but a sublinear behaviour for the solid film. At low excita-
tion densities, the magnitude of the transient absorption
signal for both samples are indistinguishable within error
margins, indicating a similar yield of long-lived charge-sepa-
rated states. However, at higher laser intensities, the ampli-
tude of the signal for the film data starts to become saturat-
ed, indicating a lower yield of such states.
Figure 6. Top: Comparison of the electron recombination dynamics at
high (black line) and low (grey line) excitation density for a solution (in
dichloromethane) and film of dyad 1. The signals have been normalised
to clearly exhibit the dynamics (t=0.1 ms). Bottom: Laser-power depend-
ence of the transient absorption signal for dyad 1 (lprobe =830 and lex
355 nm); grey line=film, black line=solution.
=
of t=7 and 0.7 ms, respectively). These dynamics are again
observed to be nonexponential and independent of excita-
tion density. The magnitude of the transient signal is ob-
served to exhibit a sublinear dependence upon excitation
density, both for the solution and solid film data. This be-
haviour is in contrast to that of dyad 1, in which such behav-
iour was only observed in the solid film (absolute compari-
son of the film and solution signal magnitude for this dyad
was not possible due to the scattering nature of the solid
film).
Discussion
The faster recombination dynamics observed in solution for
the more conjugated dyad (2) is consistent with the expect-
ed stronger electronic coupling through the molecular
bridge for this dyad. However, the nonexponential nature of
the dynamics is unexpected. Most plausibly, this nonexpo-
nential behaviour arises from different structural conforma-
tions of the dyad in solution, varying the electronic coupling
+
An analogous comparison of solution and film data was
undertaken for dyad 2, as shown in Figure 7. Most strikingly,
and in marked contrast to dyad 1, it is apparent that the re-
combination dynamics for the solid film are an order of
magnitude slower than those observed in solution (lifetimes
between the exTTFC and C60Cꢀ species. For dyad 1, the mag-
nitude of the transient signal exhibits the expected linear de-
pendence on excitation density, consistent with negligible
dyad–dyad interactions. However, for dyad 2, the transient
signal exhibits a sublinear dependence on excitation density,
Chem. Eur. J. 2005, 11, 7440 – 7447
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7445

E. Palomares, N. Martín et al.
cannot be attributed to bimolecular recombination process-
es. Rather, the accelerated recombination dynamics are as-
signed to stronger electronic coupling between exTTFC+ and
C60Cꢀ geminate radical pairs in the film. Such stronger cou-
pling may result from enhanced intramolecular electronic
coupling within each dyad owing, for example, to a more fa-
vourable structural conformation or to strong intermolecular
interactions between an exTTF moiety on one dyad and a
C60 moiety on a neighbouring dyad. In either case, the re-
sulting fast recombination dynamics suggest that this dyad
will not be attractive for photovoltaic device applications.
On the other hand, for dyad 2, it is striking that the recom-
bination dynamics are observed to be ten times slower in
the solid film relative to solution. This retardation is consis-
tent with a recent study of Janssen et al. of OPV–C60 dyads
in solution and solid films.[23] It is possible that the recombi-
nation dynamics for dyad 2 also originate from geminate re-
combination processes, as discussed above for dyad 1, but
with reduced rather than increased electronic coupling in
the film relative to the solution. However, both the lower di-
electric constant expected for films compared with solutions
(erꢁ3–4 in film compared with 9 in solution),[40,41]and the
more dense molecular packing in the film, suggest that the
electronic coupling for solid films should be higher rather
than lower than that in solution, contrary to this interpreta-
tion. More plausibly, the slow recombination dynamics ob-
served in the solid film can be assigned to dissociation of
the photogenerated radical pairs to free carriers, and subse-
quent trapping of these free carriers on low-energy sites
within the film. Such a model of slow, detrapping-limited re-
combination dynamics would be consistent with studies of
random polymer–C60 blends, both carried out by us and
other groups. This interpretation requires dissociation of the
geminate radical pairs, in contrast to dyad 1, where the ex-
perimental data is consistent with geminate charge-recombi-
nation dynamics, both in solution and the solid film. En-
hanced dissociation of geminate radical pairs for dyad 2 is
consistent with the expected greater p-orbital stacking of
this dyad, as evidenced by the dyad aggregation observed
even in solution. The enhanced radical-pair dissociation and
slow charge-recombination dynamics can be expected to
favour efficient photovoltaic device function, although we
note that the poor optical quality of the spun coated films in
practice limit device applications for this dyad.
Figure 7. Top: Comparison of the electron recombination dynamics at
high (black line) and low (grey line) excitation density for a solution (in
dichloromethane) and film of dyad 2. The signals have been normalised
to clearly exhibit the dynamics (t=0.1 ms). Bottom: Laser-power depen-
dence of the transient absorption signal for dyad 2 (lprobe =830 and lex
355 nm); grey line=film, black line=solution.
=
tending to saturate in magnitude as the excitation density
was increased. This behaviour was observed in both concen-
trated (ꢁ0.06 mm) and dilute (0.01 mm) solutions. This non-
linear behaviour suggests significant dyad–dyad interactions
(for example, exciton–exciton annihilation or bimolecular
recombination processes), indicative of significant dyad ag-
gregation even in dilute solution. Such aggregation is consis-
tent with the scattering nature of the solid films fabricated
with this dyad, and further supported by atomic force micro-
scopy (AFM) studies of the solid films for both dyads. We
believe that dyad 2 might show a higher degree of molecular
ordering than dyad 1 by p–p stacking between the fully con-
jugated molecular bridges. These molecular aggregates were
previously demonstrated for other systems in which the
presence of conjugated units induced ordered structures. For
this reason, the presence of molecular aggregation is tenta-
tively attributed to the more planar and conjugated nature
of dyad 2 relative to dyad 1.
Efficient photovoltaic device function requires the photo-
generation of a high yield of long-lived charge-separated
species. On the basis of solution dynamics alone, dyad 1 ap-
pears most attractive for such applications. However, it is
apparent from the experimental data reported here that
such solution data cannot be readily extrapolated to solid
films. After consideration of the charge-recombination dy-
namics observed in solid films, it is apparent that the elec-
tron-transfer dynamics for dyad 2 are in fact more suitable
for photovoltaic applications. Several factors influence the
electron-transfer dynamics in the solid film which are not
present in solution. Firstly, donor–acceptor electronic cou-
pling may be very different in solution and in solid films. A
For dyad 1, the recombination dynamics observed for the
solid film are approximately ten times faster than those ob-
served in solution. In solution, these recombination dynam-
ics are assigned to geminate, monomolecular charge recom-
+
bination of exTTFC /C60Cꢀ radical pairs in isolated dyad mol-
ecules, consistent with the observed excitation-density de-
pendence. The observed excitation-density independence of
the recombination dynamics in the solid film suggests that
the faster recombination dynamics observed for solid films
7446
www.chemeurj.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 7440 – 7447

exTTF–Bridge–C60 Dyads
FULL PAPER
[11] K. Ohkubo, H. Kotani, J. Shao, Z. Ou, K. M. Kadish, G. Li, R.
Pandey, M. Kujitsuka, O. Ito, H. Imahori, S. Fukuzumi, Angew.
Chem. 2004, 116, 871; Angew. Chem. Int. Ed. 2004, 43, 853.
[12] G. Steinberg-Yfrach, J. L. Rigaudt, E. N. Durantini, A. L. Moore, D.
Gust, Nature 1998, 392, 479.
key concern here is the potential for intermolecular interac-
tions between a donor on one molecule and an acceptor on
a neighbouring species. Long geminate radical-pair lifetimes
may therefore require consideration of not only intramolec-
ular but also intermolecular electronic interactions.
A
[13] N. Armaroli, F. Barigelletti, P. Ceroni, J. F. Eckert, J. F. Nicoud, J. F.
Nierengarten, Chem. Commun. 2000, 599.
[14] J. F. Nierengarten, J. F. Eckert, J. F. Nicoud, L. Ouali, V. Krasnikov,
G. Hadziioannou, Chem. Commun. 1999, 617.
[15] J. F. Eckert, J. F. Nicoud, J. F. Nierengarten, S. G. Liu, L. Echegoyen,
F. Barigelletti, N. Armaroli, L. Ouali, V. V. Krasnikov, G. Hadziioan-
nou, J. Am. Chem. Soc. 2000, 122, 7467.
[16] E. Peeters, P. A. van Hal, J. Knol, C. J. Brabec, N. S. Sariciftci, J. C.
Hummelen, R. A. J. Janssen, J. Phys. Chem. B 2000, 104, 10174.
[17] J. L. Segura, R. Gómez, N. Martín, C. Luo, A. Swartz, D. M. Guldi,
Chem. Commun. 2001, 707.
second consideration is the efficiency of geminate radical-
pair dissociation to yield free carriers, requiring favourable
intermolecular acceptor–acceptor and/or donor–donor inter-
actions. The success of dyad 2 relative to dyad 1 in generat-
ing long-lived charge-separated species in solid films, as re-
ported here, appears to derive primarily from this latter con-
sideration, with p-orbital stacking favouring efficient radi-
cal-pair dissociation.
[18] P. A. van Hal, R. A. J. Janssen, G. Lanzani, G. Cerullo, M. Zavelani-
Rossi, S. De Silvestri, Phys. Rev. B 2001, 64, 075206.
[19] N. Armaroli, G. Accorsi, J.-P. Gisselbrecht, M. Gross, V. Krasnicov,
D. Tsamouras, G. Hadziioannou, M. J. Gómez-Escalonilla, F. Langa,
J. F. Eckert, J. F. Nierengarten, J. Mater. Chem. 2002, 12, 2077.
[20] J. F. Nierengarten, N. Armaroli, G. Accorsi, Y. Rio, J. F. Eckert,
Chem. Eur. J. 2003, 9, 37.
[21] T. Gu, J. F. Nierengarten, Tetrahedron Lett. 2001, 42, 3175.
[22] T. Gu, D. Tsamouras, C. Melzer, V. Krasnikov, J.-P. Gisselbrecht, M.
Gross, G. Hadziioannou, J. F. Nierengarten, ChemPhysChem 2002,
3, 124.
[23] P. A. van Hal, S. C. J. Meskers, R. A. J. Janssen, Appl. Phys. A 2004,
79, 41.
[24] N. Martín, L. Sánchez, C. Seoane, E. Ortí, P. M. Viruela, R. Viruela,
J. Org. Chem. 1998, 63, 1268.
[25] Fundamentals and Applications of Tetrathiafulvalene, Vol. 16 (Eds.:
J. Yamada, T. Sugimoto), Springer, Berlin, 2004.
[26] N. Martín, L. Sánchez, D. M. Guldi, Chem. Commun. 2000, 113.
[27] M. A. Herranz, B. Illescas, N. Martín, C. Luo, D. M. Guldi, J. Org.
Chem. 2000, 65, 5728.
Conclusion
The application of molecular donor–acceptor structures to
photovoltaic energy conversion requires a detailed under-
standing of the correlations between molecular structure,
electron-transfer dynamics and device function. More de-
tailed studies, including different temperatures during thin-
film formation, will be necessary to completely understand
the charge-transfer process at interfaces in organic molecu-
lar devices. It can be concluded from the studies reported
here that transient absorption studies of solid molecular
films is a powerful approach to elucidating such correlations,
and may play an effective role in guiding future optimisation
of molecular structure for such applications.
[28] D. M. Guldi, L. Sánchez, N. Martín, J. Phys. Chem. B 2001, 105,
7139.
Acknowledgements
[29] M. A. Herranz, N. Martín, J. Ramey, D. M. Guldi, Chem. Commun.
2002, 2968.
[30] S. Gonzalez, N. Martín, D. M. Guldi, J. Org. Chem. 2003, 68, 779.
[31] S. González, N. Martín, A. Swartz, D. M. Guldi, Org. Lett. 2003, 5,
557.
[32] M. C. Díaz, M. A. Herranz, B. M. Illescas, N. Martín, N. Godbert,
M. R. Bryce, C. Luo, A. Swartz, G. Anderson, D. M. Guldi, J. Org.
Chem. 2003, 68, 7711.
We would like to acknowledge financial support from BP Solar (British
Petroleum), the Engineering and Physical Sciences Research Council
(EPSRC) and the Ministerio de Ciencia y Tecnología of Spain (project
BQU2002-00855). E.P. would like to acknowledge the EU for the Marie
Curie Fellowship HPMF-CT-2002-01744 for financial support and the
MEC for the Ramon y Cajal Fellowship.
[33] M. C. Díaz, B. M. Illescas, N. Martín, P. M. Viruela, R. Viruela, E.
Ortí, O. Brede, I. Zilbermann, D. M. Guldi, Chem. Eur. J. 2004, 10,
2067.
[34] D. M. Guldi, G. de la Torre, F. Giacalone, J. L. Segura, N. Martín,
Chem. Eur. J. 2005, 44, 1267.
[1] F. Padinger, R. S. Rittberger, N. S. Sariciftci, Adv. Funct. Mater.
2003, 13, 85.
[2] M. R. Wasielewski, Chem. Rev. 1992, 92, 435.
[35] L. Sánchez, I. Pérez, N. Martín, D. M. Guldi, Chem. Eur. J. 2003, 9,
2457.
[36] D. M. Guldi, C. Luo, A. Swartz, R. Gómez, J. L. Segura, N. Martín,
J. Phys. Chem. B 2004, 108, 455.
[3] D. Gust, T. A. Moore, A. L. Moore, Acc. Chem. Res. 2001, 34, 40.
[4] H. Imahori, Y. Mori, J. Photochem. Photobiol. C 2003, 4, 51.
[5] R. Holzwarth, M. Katterle, M. G. Muller, Y. Ma, V. Prokhorenko,
Pure Appl. Chem. 2001, 73, 469.
[37] U. Stalmach, H. Kolshorn, I. Brehm, H. Meier, Liebigs Ann. 1996,
1449.
[38] D. M. Guldi, C. Luo, A. Swartz, R. Gómez, J. L. Segura, N. Martín,
C. J. Brabec, N. S. Sariciftci, J. Org. Chem. 2002, 67, 1141.
[39] M. Maggini, M. Prato, G. Scorrano, J. Am. Chem. Soc. 1993, 115,
9798.
[40] J. A. Riddick, W. B. Bunger, Organic Solvents, Wiley, New York,
1970.
[41] K. R. Seddon, A. Stark, M. J. Torres, Pure Appl. Chem. 2000, 72,
2275.
[6] P. Samorì, X. Yin, N. Tcheborateva, Z. Wang, T. Pakula, F. Jackel,
M. D. Watson, A. Venturini, K. Mullen, J. P. Rabe, J. Am. Chem.
Soc. 2004, 126, 3567.
[7] a) K. G. Thomas, V. Biju, P. V. Kamat, M. V. George, D. M. Guldi,
ChemPhysChem 2003, 4, 1299; b) F. Giacalone, J. L. Segura, N.
Martín, D. M. Guldi, J. Am. Chem. Soc. 2004, 126, 5340.
[8] a) I. B. Martini, B. Ma, T. Da Ros, R. Hegelson, F. Wudl, B. J.
Schwartz, Chem. Phys. Lett. 2000, 327, 253; b) G. De la Torre, F.
Giacalone, J. L. Segura, N. Martín, D. M. Guldi, Chem. Eur. J. 2005,
11, 1267.
[9] A. Harriman, Angew. Chem. 2004, 116, 5093; Angew. Chem. Int. Ed.
2004, 43, 4985.
[10] S. C. J. Meskers, P. A. van Hal, A. J. H. Spiering, A. F. G. van der
Meer, J. C. Hummelen, R. A. J. Janssen, Phys. Rev. B 2000, 61, 9917.
Received: December 20, 2004
Revised: August 5, 2005
Published online: October 27, 2005
Chem. Eur. J. 2005, 11, 7440 – 7447
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7447