Synthesis and photophysical properties of ferrocene-oligo(benzoateethynylene)- fulleropyrrolidines dyads and triads. Implications in photovoltaic cells

Article abstract of DOI:10.1016/j.jorganchem.2018.02.040

A series of fulleropyrrolidines-conjugated bridge-ferrocene or triazene oligomers were selectively synthesized by the Sonogashira reaction by applying the step-by-step approach. The bridge is constituted by 1, 2 and 3 benzoateethynylene units (BzE) and bears triazene (Et₂N₃) or ferrocene as terminal groups affording the C60-2PEN₃ and C60-3PEN₃ dyads and C60-1PEFe, C60-2PEFe and C60-3PEFe triads. DQF-COSY, HETCOR, ¹H and ¹³C NMR and the MALDI-TOF characterization clearly confirmed the expected molecular structure. The absorption spectra of the fulleropyrrolidine oligomers do not match the sum of the individual spectra: N-methylfulleropyrrolidine (NMF) and BzEs, suggesting electronic interaction between the two moieties in the ground state. The fluorescence of the BzE is strongly quenched after functionalization with NMF, which could be indicative of energy or electron transfer from the triazene or ferrocene as electron donor to the fulleropyrrolidine electron acceptor through the π-bridge. The latter process was confirmed by cyclic voltammetry. The strength of the electron-accepting group gets to increase anodically the oxidation potential, or decrease cathodically the reduction potential in the order C60-pyrrolidine > benzoate. The character of the HOMO in the series is defined by the electron-donating ferrocene or triazene moiety, whereas the character of the LUMO is mainly determined by the electron-accepting group and is further supported by theoretical calculations. Photovoltaic devices presented low efficiencies, due to the absorption range of the oligomers being out of the maximum solar irradiance and the inhomogeneous organization in the films.

Full text of DOI:10.1016/j.jorganchem.2018.02.040

Accepted Manuscript
Synthesis and photophysical properties of ferrocene-oligo(benzoateethynylene)-
fulleropyrrolidines dyads and triads. Implications in photovoltaic cells
J.R. Flores, F. Perez, R.M. Jimenez-Barrera, E. Arias, I. Moggio, R. Torres, G.
Rodríguez, M. Ottonelli, R.F. Ziolo
PII:
S0022-328X(18)30139-6
10.1016/j.jorganchem.2018.02.040
JOM 20339
DOI:
Reference:
To appear in: Journal of Organometallic Chemistry
Received Date: 21 November 2017
Revised Date: 21 February 2018
Accepted Date: 23 February 2018
Please cite this article as:
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Synthesis and Photophysical Properties of Ferrocene- oligo(benzoateethynylene)-
fulleropyrrolidines Dyads and Triads. Implications in Photovoltaic Cells
J.R. Flores1, F. Perez1, R. M. Jimenez-Barrera1, E. Arias*1, I. Moggio*1, R. Torres1,
G. Rodríguez1, M. Ottonelli2, R. F. Ziolo1
1Centro de Investigación en Química Aplicada, Boulevard Enrique Reyna 140, 25253
Saltillo, Coahuila, México. Tel: +52 844 438 9830, eduardo.arias@ciqa.edu.mx,
ivana.moggio@ciqa.edu.mx
2Dipartimento di Chimica e Chimica Industriale, Università degli Studi di Genova, Via
Dodecaneso 31 I-16146, Genova, Italia

ACCEPTED MANUSCRIPT
Abstract
A series of fulleropyrrolidines-conjugated bridge-ferrocene or triazene oligomers were
selectively synthesized by the Sonogashira reaction by applying the step-by-step
approach. The bridge is constituted by 1, 2 and 3 benzoateethynylene units (BzE) and
bears triazene (Et2N3) or ferrocene as terminal groups affording the C60-2PEN3 and
C60-3PEN3 dyads and C60-1PEFe, C60-2PEFe and C60-3PEFe triads. DQF-COSY,
1
13
HETCOR, H and C NMR and the MALDI-TOF characterization clearly confirmed the
expected molecular structure. The absorption spectra of the fulleropyrrolidine oligomers
do not match the sum of the individual spectra: N-methylfulleropyrrolidine (NMF) and
BzEs, suggesting electronic interaction between the two moieties in the ground state.
The fluorescence of the BzE is strongly quenched after functionalization with NMF,
which could be indicative of energy or electron transfer from the triazene or ferrocene as
electron donor to the fulleropyrrolidine electron acceptor through the π-bridge. The latter
process was confirmed by cyclic voltammetry. The strength of the electron-accepting
group gets to increase anodically the oxidation potential, or decrease cathodically the
reduction potential in the order C60-pyrrolidine>benzoate. The character of the HOMO
in the series is defined by the electron-donating ferrocene or triazene moiety, whereas
the character of the LUMO is mainly determined by the electron-accepting group and is
further supported by theoretical calculations. Photovoltaic devices presented low
efficiencies, due to the absorption range of the oligomers being out of the maximum
solar irradiance and the inhomogeneous organization in the films.
Keywords: Ferrocene, energy transfer, fullerenes, HOMO-LUMO levels
1. Introduction
Organic photovoltaic technology (OPV) has emerged since the discovery that

ACCEPTED MANUSCRIPT
photoinduced electron transfer process could be achieved from fullerene as an electron
acceptor to conjugated polymers as electron donor materials [1]. The main advantages
of OPVs were highlighted in the earliest reports and included the possibility to
chemically modulate the energy levels of the active materials to get an efficient
photoinduced mechanism, and to process larger individual cells and on any substrate
geometry through low cost techniques. Encouraged by these expected advantages,
many groups around the world have been searching for both, the most efficient
photovoltaic macromolecule and the optimal device configuration. Despite that, and
though great improvements have been made in both lines, typical efficiencies for OPV
cells are only around 8-10 % in single junction and with a recent breakthrough of 12.7%
in a tandem multichromophoric two-subunit configuration [2]. Generally, the
photoinduced process is obtained in a so-called heterojunction configuration, where the
electron donor (D) and acceptor (A) materials are physically mixed as the active layer.
This strategy is usually more practical as it opens the way to chemically synthesize
different D and A materials in order to modulate the optical and electrical properties [3].
However, the photovoltaic performance of the cell is strongly morphology dependent. An
interesting approach is the “molecular cable” or “push-pull” system, where D and A are
covalently linked units (usually through a π-transport bridge). Although this strategy
implies, in general, larger and more complex synthetic work, it is expected to reduce the
problems related to the phase segregation and morphology as D/A are chemically
bonded. Moreover, as the feasibility of obtaining a photoinduced process can be
estimated through the Weller equation [4], the molecular cable materials represent ideal
models for understanding the photophysics of organic photovoltaic devices. Among the
organic A materials, even though graphene [5], carbon nanotubes [6], and perylenes [7]
have been proposed, fullerenes are still the most used [8]. In the literature, a large
number of scientific reports on the synthesis of fullerene-oligo(phenylenevinylene)s
molecular cable materials (dyads [9], triads [10,11] and more complex systems [12] can
be found).
The state of art of fullerene-phenylenethynylene (PE) derivatives, however, is more
restricted [13], likely due to the fact that the rigidity of PEs promotes time dependent
aggregation, and in agreement with Martin et al. [13b] phenyleneethynylenes present

ACCEPTED MANUSCRIPT
(PPE) polarization of the triple bonds, their shorter bonds lengths relative molecules with
double bonds make that the electron transfer pathway through the PPE bridges is
interrupted by the ethynylene bridge, influencing the charge-separation process and
explaining the difference in electron-transfer properties between PPEs and PPVs, but
the photovoltaic properties of PPEs materials is relatively new.
In this work, we report on the synthesis of a series of oligomers as precursors and
models of the oligo(fulleropyrrolidines) with the aim to provide new push pull models for
studying intrinsic photovoltaic process in organic materials. The series provides
molecules of increasing length where the ferrocene or triazene as the electron-donor
group connects to C60-pyrrolidine as the electron-accepting group in a configuration
D→π→A. The photophysical study reveals for the two C60-2PEN3 and C60-3PEN3
named as dyads and the three C60-1PEFe, C60-2PEFe and C60-3PEFe named as
triads, an exergonic photoinduced electron transfer from the ferrocene or triazene to
C60-pyrrolidine, upon excitation of either the BzE unit (likely preceded by energy
transfer) or the fullerene. This process is also confirmed by cyclic voltammetry, where
the electron transfer is achieved from the iron centre of the ferrocene to the electron-
accepting group through the π-bridge. The strength of the electron-accepting group gets
to increase anodically the oxidation potential, or decrease cathodically the reduction
potential. Despite that, at least in solution, the photoinduced process falls in the inverted
Marcus region, i.e., charge recombination is expected to occur, which could explain the
low efficiencies (η) found in “molecular cable” solar cells. Improvements in η are
obtained by using these materials as electron acceptors in substitution of PC60BM in
heterojunction solar cells based on poly(3-hexyltiophene) (P3HT). This result suggests
that the new materials do not have a balance in the charges, likely because the BzE unit
acts not only as a transport bridge but also as electron withdrawing material because of
the ester group. In this context, it is however important to mention that the solar cells
were fabricated and studied at ambient conditions and under AM1.5 solar simulator
irradiation. Higher efficiencies are expected if devices are prepared and analysed in
controlled atmosphere and upon selective irradiation (with a UV laser) as reported in
other works on cable materials. This work opens the way to design polar C60-PE
molecular cable materials with BzE units substituted with, for instance, glycol lateral

ACCEPTED MANUSCRIPT
chains that could be deposited on active layers from polar solvents e.g. P3HT, PTB7 to
form multilayer tandem photovoltaic devices.
2. Results and discussion
2.1 Syntheses and chemical characterization. Prior to the oligomer synthesis, a
bifunctional monomer 4 was prepared in three steps, scheme 1. Esterification of 1 was
carried out with (9-bromononyloxy)(tert-butyl)dimethylsilane by using DBU in toluene at
reflux to give 2, which was Pd cross-coupled with TMSA to obtain 3 in 91 % yield.
Desilylation of 3 with –F+NBu4 in THF affords the desired monomer 4 in 90 % yields.
Scheme 1. Reagents and conditions: (a) DBU, toluene, reflux, 16h; (b) [(C6H5)3P]2PdCl2 (3.0 %
mol), CuI (1.5 % mol), TMSA, Et3N, 60°C, 16h; (c) –F+NBu4, THF, 0°C, 30 min.
The pathway to synthesize the aldehyde-terminated oligomers is described in scheme
2. The convergent-divergent synthesis in general involves three repetitive reactions at
each cycle: Sonogashira Pd/Cu cross coupling reaction, iodination and reduction of the
aldehyde group.
Scheme 2. Reagents and conditions: (d) [(C6H5)3P]2PdCl2 (3.0 % mol), CuI (1.5 % mol), 60°C,
16h; (e) NaBH4, THF, N2, 50°C, 2h; (f) CH3I, 120°C vacuum, 15h.

ACCEPTED MANUSCRIPT
The oligomerization begins with the Pd/Cu cross coupling of 4-bromobenzaldehyde 5
with commercial ethynylferrocene 6 to generate 7, which one part is used for the C60
cycloaddition and the other part is subjected to the sodium borohydride reduction to
obtain the benzyl-alcohol 8, HO-PEFe. It was found that 4 eq. of NaBH4 assure the
aldehyde reduction in almost 85-90% yield, however yields also depend on both of the
next actions: bubbling the solution with nitrogen before the reaction and not allowing the
temperature to pass 60 °C. By Pd/Cu cross-coupling 5 with 4, the aldehyde-triazene
terminated oligomer 9 is obtained. At this point 9 is divided into two portions; one is
subjected to the NaBH4 reduction to generate the ethyl alcohol 10, HO-2PEN3, while
with the other portion, the triazene is converted to iodine group with iodomethane in a
sealed capsule under vacuum. For this reaction, we found that the yield to obtain 11
depends on the precise temperature control at 120°C ± 5°C and on the vigorous stirring
of the reaction medium. Then 11 is divided into two portions; one undergoes Pd/Cu
cross-coupling with ethynylferrocene to generate 12 (which in turn, one part is used for
the C60 cycloaddition), while the other part is subjected to the NaBH4 reduction to obtain
the ethyl-alcohol 13, HO-2PEFe. The second portion of 11 is subjected to another
divergent cycle, by carrying out a Pd/Cu cross-coupling reaction with 4, affording the
aldehyde-triazene terminated oligomer 14, in which one part is used for the C60
cycloaddition and the other for the NaBH4 reduction, generating the ethyl alcohol 15,
HO-3PEN3. The triazene conversion of 14 to iodine group in 16 allows its subsequent
Pd/Cu cross coupling with ethynylferrocene, forming the aldehyde-ferrocene terminated
oligomer 17; which in turn, one portion is used for the C60 cycloaddition, while the other
is subjected to the NaBH4 reduction reaction generating the ethyl alcohol 18, HO-
3PEFe. As can be seen, following these repetitive reactions cycles, oligomers can be
grown, until they reach the limit of their solubility.
The synthesis route of the fullerene derivatives by the Prato’s [14] reaction is depicted in
scheme 3. The [3+2] 1,3-dipolar cycloaddition of azomethine ylides generated from the
aldehydes 7, 9, 12, 14, 17 with the sarcosine to the C60 gave, in general, yields ranging
from 35 to 52 %, as reported in literature [14]. It is worth mentioning that the byproducts

ACCEPTED MANUSCRIPT
from the reaction consisted of the precursor oligomers but without the aldehyde group.
Two series of C60-PEs were synthesized by varying both the oligomer length and the
type of termini group: those bearing triazene as terminal group, dyads: C60-2PEN3 and
C60-3PEN3 and those bearing the ferrocene, tryads: C60-1PEFe, C60-2PEFe and C60-
3PEFe. All of the fullerene derivatives are soluble in CH2Cl2, CHCl3, toluene, THF and
are insoluble in methanol, acetone and acetonitrile. In our original idea, the OH of the 9-
nonecanol benzoate side chains was intended to serve as a hydrophilic group that could
allow the Langmuir-Blodgett films formation; however, at the moment of cleavage of the
tert-butyldimethylsilane of the -OH protecting group, the derivative C60-3PEFe became
insoluble overtime. In addition, we observed that all of the ferrocene compounds with
and without fulleropyrrolidines in solution are prone to oxidize overtime, likely because
ferrocene is not only a good electron donor but an excellent light trapping group that
may promote oxidation (vide infra). The oxidation process is evidenced because: i)
insoluble aggregates are observed, and ii) their 1H NMR spectra develop paramagnetic
broad resonant bands instead of the well resolved resonant signals, see Fig 1Sa as
example. However, by passing the solution through a plug of alumina, molecules
recover their original well-resolved resonant signals, but with at least a 5 % loss of
material with each pass. The TGA revealed that fulleropyrrolidines are stable up to
200°C, Fig 1Sb.
Scheme 3. Reagents and conditions: (g) aldehyde (1 eq), C60 (1 eq), sarcosine (2.63 eq),
toluene (75 mL/100 mg of aldehyde) at reflux, 18h.
The cycloaddition reaction of aldehydes 7, 9, 12, 14 and 17 to the C60 is confirmed

ACCEPTED MANUSCRIPT
because the singlet signal of the aldehydes (~10 ppm) disappears, Fig 4S-5S (ESI)
giving rise to the resonance of the pyrrolidine protons, such as that of the –NCH3 group,
which clearly appears at 2.8 ppm. As an example of the series, the spectrum of C60-
3PEFe compound is shown in Figure 1, and the rest of the oligo(fulleropyrrolidines) are
shown in Fig 7S. The singlet of the C(2)H-pyrrolidine is centered at 4.95 ppm, while one
3
of the germinal C(1)H1 protons appears as a doublet at 4.96 ppm with a JH1-H2 = 9.6
Hz, the other C(1)H2 proton resonates in a high signal density region and is not seen.
Figure 1. Comparative 1H NMR spectra of C60-3PEFe in CDCl3 + CS2 recorded in: (a) a
300 MHz equipment and (b) in a 500 MHz equipment; multiplicity of H1 and H3 protons
are better resolved in a 500 MHz equipment.
In this respect, in the expanded Figure 1b, one can identify those signals of the
monosubstituted ferrocene that appear as two triplets at 4.52 ppm (H5,5’) and at 4.28
ppm (H6,6’) due to their magnetic nonequivalence (giving the interaction of four nuclei

ACCEPTED MANUSCRIPT
3
AA’BB’ system) with a typical constant JH5-H6 of 1.65 Hz. Moreover, it is possible to
distinguish the quasi-overlapping triplet at 4.34 ppm (3JH8-H9 = 6.6 Hz) that corresponds
to the non-chemically equivalents COOCH2 and the singlet at 4.26 ppm of the Cp(H7).
The distinctive C(1)H2 proton could be unambiguously assigned by both, DQF and
COSY, where the correlation C(1)H1-H2 is clear; as reference, the -N-CH3 does not
correlate with any proton, Fig 8S, and by HETCOR, Fig 9S, where C2 correlates with
H3, and C1 with H1 and H2. In what concerns the 13C spectra (broadband decoupling),
Fig 10S, the unambiguous chemical shift assignment was achieved by a combination of
APT, DEPT-135 13C, experiments. The lack of C2 or Cs symmetry of the
oligo(fulleropyrrolidines) is helpful to differentiate for instance, the two carbonyls that
resonate at different displacement, i.e., at 165.85 ppm for C36 and 165.51 for C29. It
was also possible to differentiate the 6 ethynylene carbons e.g. for C60-3PEFe at 96.40,
95.15, 92.17, 90.5, 88.95, 84.77 ppm. Notice that C2 of the pyrrolidine resonates in the
region of the C≡C carbons at 83.34 ppm, while that of the C1 at 70.17 ppm overlaps
1
with those of the Cp(H7). In general, all of the H and 13C resonant signals of the
fulleropyrrolidines coincide with those reported in literature [15].
All of the C60-nPEN3 and C60-nPEFe series were also analyzed by MALDI-TOF mass
spectrometry, by using the α-cyano-4-hydroxy cinnamic acid (CHCA) and 2-[(2E)-3-(4-
tert-butylphenyl)-2-methylprop-2-enylidene] malonitrile (DCTB) matrices. The molecular
ions for all the oligo(fulleropyrrolidines) could be clearly observed with DCTB matrix in
the negative ion mode, while in the positive ion mode or with the CHCA matrix, the
-
molecular ion peaks were found to be very weak. For C60-2PEN3: m/z 1353.55 (M. ,
calcd for C97H56N4O3Si: 1352.41); C60-3PEN3: m/z 1753.80 (M.-, calcd for
C121H92N4O6Si2: 1752.66); C60-2PEFe: m/z 1461.32 (M.- calcd for C105H55FeNO3Si:
1461.33); C60-3PEFe: m/z 1861.55 (M.- calcd for C129H91FeNO6Si2: 1861.57).
2.2 Electrochemical properties
We used cyclic voltammetry to evaluate the donor-acceptor electronic influence of the
compounds herein synthesized. The electrochemical data are summarized in Table 1.
We observed an insoluble precipitate at the end of the redox analyses in CH2Cl2, which
confirmed NMR observations of the compounds paramagnetic behaviour once oxidized

ACCEPTED MANUSCRIPT
[16]. Structurally, the oligomers have a ferrocene, a phenyleneethynylene(s) moiety, a
benzoate(s) and or triazene as possible redox groups, while the fulleropyrrolidines in
addition have the fullerene-pyrrolidine; both systems are composed of D-(π)bridge-A
configuration.
Table 1. Electrochemical properties at the maxima redox potentials of model oligomers
and oligo(fulleropyrrolidines) synthetized and studied in this work (0.5 mmol) in CH2Cl2,
0.1 M Bu4NPF as electrolyte, at scan rate of 50 mVs-1 (V vs Ag/AgCl) and using GCE as
working electrode.
III
II
I
I
II
III
Molecule
Ered
Ered
Ered
Eox
Eox
Eox
HOMO LUMO
Eg
(V)
(V)
(V)
(V)
(V)
(V)
(eV)
(eV)
(eV)
2.04
−1.523 −1.020 -0.660 +1.386 +2.329
+0.391 +0.482
-6.18
-4.14
MFP
Ferrocene
HO-2PEN3
C60-2PEN3
HO-3PEN3
C60-3PEN3
HO-1PEFe
C60-1PEFe
HO-2PEFe
C60-2PEFe
HO-3PEFe
C60-3PEFe
-0.970 +1.144 +1.967 +2.433
-0.868 +1.166 +1.996 +2.430
-0.860 +0.950 +1.983 +2.350
-0.828 +1.128 +1.980 +2.300
-5.94
-5.96
-5.75
-5.92
-5.61
-5.66
-5.40
-5.43
-5.34
-5.37
-2.83
-3.93
-3.94
-3.97
-3.41
-3.58
-3.52
-3.72
-3.77
-3.84
3.10
2.03
1.81
1.95
2.20
2.09
1.88
1.71
1.57
1.54
-1.065
-1.165
-1.048
-1.390 +0.816
-1.220 +0.868
+1.829
+1.681
−1.682 -1.271 +0.609 +0.884 +1.697
-1.081 +0.635 +0.903 +1.644
-1.025 +0.546 +0.829 +1.613
-0.962 +0.575 +0.866 +1.640
As a term of comparison, Figure 2 shows the voltammograms of some representative
oligomers. In general, we observe in the anodic region, waves originated by the
ferrocene, the ester that is overlapped with that of C60 and of the phenyleneethylene
groups, while in the cathodic region, just one or maximum two waves were seen. As
reference for our study, the calculated ferrocene potentials were of Eox = 0.480 V and
Ered = 0.391 V (formal potential E° = 0.436 V). The first oxidation potential is thus
assigned to the ferrocene, which shifts to higher potentials with respect to the shortest
of the oligomer HO-1PEFe (0.816 V), however by increasing the conjugation to HO-

ACCEPTED MANUSCRIPT
2PEFe (0.609 V) and to HO-3PEFe (0.546 V) a decrease in the oxidation potential is
observed, but still with higher values with respect to the free ferrocene. The same trend
is observed in the triazene series, but triazene shows to be more difficult to be oxidized
and reduced, than ferrocene. The ferrocene oligomers behaviour provides evidence that
there is an electron transfer between the iron centre of the ferrocene and the electron-
accepting benzoate(s) group and, that in general, phenyleneethynylene’s are also
difficult to be oxidized (~1.3 V) or reduced (~2.5 V) [17].
Figure 2. Comparative voltammograms of MFP, C60-3PEN3, HO-3PEFe and C60-
3PEFe representative molecules studied in this work in CH2Cl2, 0.1 M of Bu4NPF6 as
electrolyte, at scan rate of 50 mV/s and using a GCE as working electrode.

ACCEPTED MANUSCRIPT
In contrast, the decrease in the potential with respect to the elongation of the
conjugated backbone suggests that there is stabilization of the oxidized species as the
conjugated chain is enlarged, as observed for ferrocenyl-phenylenevinylene oligomers
[18]. This anodic shift is particularly significant when going from the oligomers to the
oligo(fulleropyrrolidines). In general terms, we observed that the C60-pyrrolidine
generates an increase of the oxidation potential of the wave assigned to the ferrocene,
suggesting that C60 is a stronger electron-accepting group than benzoates, as can be
seen from the values in Table 1. It is well known in cyclic voltammetry that the wave
potential gives a direct indication of the degree of the electron density transfer, which in
this case, is from the ferrocene to the acceptor group; here, the ferrocene gives a higher
electron density and will be more difficult to be oxidized, generating a shift to higher
potentials. Regarding the cathodic region, we observe only one reduction wave for
ferrocene molecules or two reduction waves for the triazene molecules. In general, we
assume that the interaction of ferrocene with the electron-accepting groups is strong for
the whole series, likely, via dynamic electron transfer that could explain why the
distance between the first redox waves is too large with respect to that of the free
ferrocene. However, this potential wave is also very wide; we infer that this is because
the wave mixes with the signal of each individual reduction group, which are certainly in
the whole system being affected by the length and type of the π-bridge and of the
electron-accepting type group. In this respect, we observed a similar trend to that of the
oxidation potential wave, but in the inverse sense. Thus, the shortest oligomer HO-
1PEFe exhibits a reduction potential wave (Ered) of -1.390 V, which decreases by
increasing the conjugation to -1.271 V for HO-2PEFe and to -1.025 V for HO-3PEFe. In
addition, the strength of the electron acceptor group makes these values decrease even
more in the order C60-pyrrolidines < benzoates. Based on these results, we can deduce
that the character of the HOMO in the series is defined by the ferrocenyl and PE moiety,
whereas the character of the LUMO is mainly determined by the electron-accepting
group, i.e., C60-fulleropyrrolidine. This was further confirmed by the theoretical
simulation of the HOMO and LUMO. Figure 3 reports the frontier orbitals for C60-2PEN3
and C60-1PEFe as representative molecules. It can be observed that the electron
density is mainly concentrated in the spatial region of the π-bridge up to triazene or

ACCEPTED MANUSCRIPT
ferrocene in the HOMOs, while it is localized on the N-methyfulleropyrrolidine in the
LUMOs, suggesting strong electronic coupling between the ferrocene or triazene and
C60-pyrrolidine. The calculated energies are in good agreement with the experimental
values obtained by cyclic voltammetry. For C60-2PEN3, HOMOth= -5.85 eV vs.
HOMOexp= -5.96 eV and LUMOth= -3.52 eV vs LUMOexp= -3.93 eV. For C60-1PEFe,
HOMOth= -5.61 eV vs HOMOexp= -5.66 eV and LUMOth= -3.54 eV vs LUMOexp= -
3.58 eV.
Figure 3. Frontier molecular orbitals; spatial distribution for the HOMO→LUMO energy levels
for: Left C60-2PEN3 and right C60-1PEFe, obtained at the B3LYP level with the 6-311+G**
and the mixed 6-311+G**(C,N,O,H)/m6-31G*(Fe) basis sets, respectively. The surfaces are
calculated at the isovalue of 130 e·nm-3 (0.02au).
2.3 Photophysical properties
2.2.1 Triazene terminated dyads
Figure 4 presents the absorption spectra of the C60-nPEN3 derivatives in CH2Cl2 and
Table 2 resume all the determined photophysical parameters. The UV-Vis spectra of N-
methylfulleropyrrolidyne (MFP) and the HO-nPEN3 oligomers are also included for sake
of discussion. The spectrum of MFP is similar to that in toluene, previously reported
[19]. It presents strong absorptions in the UV at 255 and 327 nm, which for C60 in n-

ACCEPTED MANUSCRIPT
hexanes have been ascribed to the electronic transitions from the ground state to the
61T1u and 31T1u excited states, respectively [20]. In the visible region (insert, Figure 4),
MFP shows a broad absorption with sharp peaks at 430 nm and 702 nm that were
assigned to the lowest allowed and the lowest singlet transitions in similar
fulleropyrrolidines [9c]. The benzoatheethynylene oligomers show a main absorption
peak at 350 nm and 372 nm for HO-2PEN3 and HO-3PEN3, respectively, with increasing
absorption coefficient when passing from the dimer to the trimer, in agreement with the
extension of the conjugated chain and thus the larger number of chromophores.
6
4
420 500 580 660 740
Wavelength (nm)
2
0
250 300 350 400 450 500 550 600 650
Wavelength (nm)
Figure 4. Absorption spectra of HO-2PEN3 (black filled triangles), MFP (blue crosses),
HO-3PEN3 (red filled circles), C60-2PEN3 (white filled triangles) and C60-3PEN3 (white
filled circles) in CH2Cl2. The inset shows the spectra in the 420-740 nm range
A shoulder can be clearly observed at 320 nm for HO-3PEN3, while it is just a slight
deflexion for HO-2PEN3. The fluorescence spectra are identical if the excitation
wavelength is on the main peak or the shoulder and present a strong peak at 406 nm
for HO-2PEN3 and 416 for HO-3PEN3 (Figure 5a). The emission spectrum of the latter
molecule is more resolved than that of HO-2PEN3 exhibiting excitonic features. The
fluorescence quantum yield is low because of the quenching effect of the triazene. The
excitation spectrum of both molecules (Figure 5b) presents a peak at around 300 nm
and a red shifted peak centred at 352 nm for HO-2PEN3 and 381 nm for HO-3PEN3,

ACCEPTED MANUSCRIPT
which does not change by choosing as the emission wavelength, either the main
fluorescence peak or its shoulder at 438 nm. A first comparison with the absorption
spectrum suggests that a good match exists for only the lower energy peak. However,
after deconvolution of the UV-Vis spectra (Figure 5b for HO-2PEN3 as example), it
appears a better match between the absorption and excitation spectra for both peaks.
Table 2. Photophysical properties of HO-nPEN3 oligomer models and of C60-nPEN3
dyads in CH2Cl2.
Molecule
λabs (nm)a
ε*10-4
(M-1cm-1)b
2.4
λemis
(nm)
395
Φ
(%)
1.10
Q
HO-2PEN3
HO-3PEN3
C60-2PEN3
C60-3PEN3
350
5.7
5.1
6.6
416
371
373
0.20
0.16
0.03
320 (sh), 372
320, 430, 699
320, 367, 700
6.9
6.7
amaximum wavelength (bold and underlined character), bat the maximum wavelength
The fact that the fluorescence spectra do not change with excitation wavelength and
that the excitation well match the absorption spectra suggest that there is only one
emitting species.
The UV-Vis spectra of the C60 derivatives, composed by the absorptions of the
corresponding benzoatheethynylenes and MFP do not correspond to the sum of the
individual spectra, Fig 11S. Similar observations were made by different authors on
other C60 derivatives of conjugated oligomers ascribing this behaviour to electronic
perturbation between the individual chromophores in the ground state [12d,13d,21].
This effect is more marked for the less conjugated system (C60-2PEN3).

ACCEPTED MANUSCRIPT
Si
O
C9H18
O
O
a)
b)
O
N
OH
N
Si
O
C9H18
O
N
O
N
OH
N
N
Si
O
C9H18
O
N
O
N
OH
N
Si O
C9H18
O
250 300 350 400 450 500
λ
360 410 460 510 560(nm)
Figure 5. a) Fluorescence spectra of HO-2PEN3 (black squares) and HO-3PEN3 (red
triangles) in CH2Cl2. b) Excitation spectrum (blue dotted line), absorption spectrum
(black solid line) of HO-2PEN3 in CH2Cl2 and its corresponding deconvolution absorption
bands (green squares).
The maximum wavelength red shifts from C60-2PEN3 to C60-3PEN3 due to the increase
of the central conjugated length moiety, but blue shifts with respect to the corresponding
benzoatheethynylene HO-nPEN3 precursors, which is in agreement with the electron
withdrawing effect of fullerene. The fluorescence spectra of both fullerene derivatives,
Fig 13S, are very similar and exhibit excitonic features. The maximum presents a very
low bathochromic shift with the increase in the phenylenethynylene segment and a
hypsochromic shift with respect to the oligomer precursors. The fullerene emission was
not detected, probably because i) of the restricted detection limit of our equipment, ii) it
overlaps with the replica at twice the wavelength of the benzoatheethynylene emission,
and iii) it was not possible to selectively excite the fullerene moiety. In fact, the UV-
bands overlap with the BzE absorptions and the visible absorption bands present a very
low absorption coefficient. Attempts to adjust the optical density at 0.1 for the fullerene
absorption at 430 nm or 702 nm failed, as the molecules tend to agglomerate at higher
concentration. The BzE fluorescence quantum yield, however, decreases after the MFP
functionalization, suggesting a possible photoinduced energy transfer, even if Förster

ACCEPTED MANUSCRIPT
energy transfer cannot be discarded as the emission and absorption spectra overlap
(see energy considerations vide infra).
2.2.1 C60-nPEFe triads
The photophysical properties of the fullerene ferrocene triads (C60-nPEFe) in CH2Cl2
are reported in Table 3 together with those of the corresponding BzE oligomers (HO-
nPEFe). The absorption spectra of both families (benzoatheethynylene and C60
derivatives) in CH2Cl2 are shown in Figure 6.
8
6
420 500 580 660 740
4
Wavelength (nm)
2
0
250 300 350 400 450 500 550 600 650
Wavelength (nm)
Figure 6. Absorption spectra of HO-1PEFe (black filled triangles), HO-2PEFe (white
filled circles), HO-3PEFe (green filled stars), C60-1PEFe (white filled triangles), C60-
2PEFe (red filled circles) and C60-3PEFe (white filled stars) in CH2Cl2. The inset shows
the spectra in the 420-740 nm range.
Both, the ferrocene oligomers and C60 derivatives present a main absorption in the UV
due to the π−π* electronic transition of the phenyleethynylene moiety and that shifts in
the series according with the increase in the conjugation, but blue shifts from the
oligomers to the C60 derivatives due to the electron withdrawing effect of fullerene (see
Table 3).

ACCEPTED MANUSCRIPT
Table 3. Photophysical properties of HO-nPEFe oligomer models and of C60-nPEFe
triads in CH2Cl2.
Molecule
λabs
(nm)a
303
ε*10-4
(M-1cm-1)b
0.53
λemis
(nm)
333
Φ
(%)
1.20
Q
HO-1PEFe
HO-2PEFe
HO-3PEFe
C60-1PEFe
4.74
5.22
3.32
376
397
337
10.40
15.00
0.30
328
356
4.0
17.3
250.0
304, 430, 455,
703
310, 430, 453,
703
355, 430, 460,
701
6.91
8.69
373
391
0.60
0.06
C60-2PEFe
C60-3PEFe
amaximum wavelength (bold and underlined character), bat the maximum wavelength
The spectra of ferrocene and N methylfulleropyridine are reported in Fig 12S together
with the sum spectra for each compound. From these figures, it can be observed that
the absorption behavior of the ferrocene fullerene triads is similar to that previously
discussed for the triazene series, where the absorptions of the individual conforming
chromophores, i.e., phenyleneethynylene, ferrocene and MFP, can be visualized, but
the spectra do not match the sum of their components.
In the visible, an absorption ranging from 440 to 700 (insert Figure 6) is also observed
for the oligomers and is assigned to ferrocene transitions [22]. Concerning the fullerene
derivatives, the C60 transition at 430 nm overlaps, giving a neat peak at this
wavelength, in addition the fullerene electronic transitions at 255 and 704 nm are also
present. From the absorption coefficient of C60-3PEFe (Figure 6) and MFP (Figure 4),
we can deduce an increase of the optical density at 255 nm of 64%, which is a large
antenna effect of the oligomer.
In fluorescence, the maximum red shifts along the series, as observed in the absorption
spectra, Table 3. Figure 7 shows the excitation, absorption and emission spectra for
C60-3PEFe and its precursor HO-3PEFe as examples. The emission spectra present
excitonic features, where the maximum blue shifts for the fullerene derivatives are in
agreement with the electron withdrawing effect of this group. The excitation spectra

ACCEPTED MANUSCRIPT
match well the PE absorption band at around 350 nm. The ferrocene and fullerene (for
C60-3PEFe) peaks are not observed as these chromophores do not participate in the
emission. It is worth mentioning that selective excitation of C60 was not possible due to
the overlapping of its absorption with that of the BzE segment. In addition, no
fluorescence assignable to fullerene could be detected as in the previous series.
However, also in this case, the fluorescence quantum yield strongly decreases with a
fluorescence quenching (Q) up to 250 (Table 3), likely indicative of a photoinduced
electron transfer process, such as has been reported in other works for similar
compounds [9,23,24]. However, energy transfer cannot be discarded as spectral overlap
also occurs for this series.
Figure 7. Absorption (solid line), excitation (dashed line) and fluorescence (dotted line)
of HO-3PEFe and C60-3PEFe (inserted figure) in CH2Cl2.
2.3. Energy considerations
In order to provide more insight about the fluorescence quenching, energy
considerations were realized for C60-3PEN3 and C60-3PEFc as examples of the two
series, Table 4.
Table 4. Photoinduced electron transfer parameters and calculated free energy ∆G of
C60 dyads and triads in CH2Cl2.

ACCEPTED MANUSCRIPT
Molecule
RCC
(Å)
r+
E00,BzE
GCS
kr
knr
kET
∆GBzE
(eV)
∆GC60-
λ
(ns-1) (ns-1) (ns-1)
pyrolidine
(Å)
(eV)
(eV)
(eV)
13.9
17.8
5.48
9.28
3.05
3.24
-1.52
-1.90
1.53
1.34
-0.19
0.90
0.78
0.04
0.05
1.51
8.32
5.9
C60-3PEN3
C60-3PEFe
-0.38
498.0
The values of the first excited state of the benzoateethynylene unit (BzE) (E00,BzE) was
obtained from the interception between the normalized absorption and fluorescence
spectrum. The energy of the charged state GCS was calculated by the Weller equation.
Based on these energy levels, the possible intramolecular photoinduced processes for
these two molecules in CH2Cl2 are depicted in Figure 8.
Figure 8. Energy diagram depicting the photoinduced intramolecular process for C60-
3PEN3 and C60-3PEFe in CH2Cl2.
The direct excitation of the benzoateethynylene (BzE) unit yields to the first excited state
C60-1(BE), from which the fluorescence can occur with a radiative constant kr (=φ/τ,
Table 4, Fig 14S). The fluorescence quenching is indicative of non radiative competitive
processes: energy or photoinduced charge transfer. Direct photoinduced charge
transfer processes upon excitation of both BzE and C60-pyrrolidine are exergonic
(∆G<0) because the charge state is at lower energy than both C60-1(BzE) (first excited
1
state of the BE unit) and (C60)-BzE (first excited state of C60-pyrrolidine). By
calculating the reorganization energy (λ), it was found that direct charge transfer

ACCEPTED MANUSCRIPT
process from the BzE unit always falls in the inverted Marcus [25] region (λ<-∆GBzE) i.e.
it will be followed by charge recombination. Direct charge transfer from the C60-
pyrrolidine to BzE unit presents λ>-∆GC60-pyrrolidine, i.e., after excitation of the fullerene
moiety, charge separation should occur. Direct excitation of the fullerene moiety,
however, was not experimentally attained as previously mentioned so we cannot
confirm if this process is responsible of the fluorescence quenching. Upon excitation of
the BE unit, energy transfer from C60-1(BzE) to 1(C60)-BzE is also energetically possible,
which could follow with fluorescence or phosphorescence (through intersystem
crossing) emission of C60 and/or to charge transfer. As previously mentioned, emission
of C60 was no detected even upon excitation of the BzE unit, so we infer that the
fluorescence quenching is mainly due to indirect photoinduced charge transfer: upon
excitation of the BzE unit, first energy transfer occurs from BzE to C60-pyrrolidine to
1
BzE reaching the first excited state (C60)-BzE. Then, photoinduced charge transfer
follows getting the charge state, as found for C60-OPV dyads upon excitation of the
OPV unit [9a]. The energy transfer constant kET =(Q-1)/τBE (Table 4) could be obtained
from the fluorescence quenching of the BzE unit Q and the fluorescence lifetime of the
BzE precursor, τBzE (0.97 ns and 0.50 ns for HO-3PEN3 and HO-3PEFe), respectively.
This gives a value of two orders larger for the ferrocene compound with respect to the
triazene derivative revealing a more efficient energy transfer, as first deactivation
process, for this triad.
2.6 Photovoltaic properties
From all of the series, just C60-3PEN3 and C60-3PEFe gave moderate optical quality
films on ITO, thus photovoltaic properties were tested: 1) intrinsically; as molecular
cables and 2) mixed with the P3HT electron donor material by substituting the PC61BM
as electron attractor material. The active layer was sandwiched between PEDOT:PSS
and aluminum or Field’s Metal electrodes, without significant differences from one to the
other cathode as has been reported [26]. Since, devices were tested at ambient
conditions, a P3HT:PC61BM (1:1 wt. %) device was built as a reference yielding a ηAM1.5
= 1.77 %, while in the literature under the same conditions a ηAM1.5 = 1.54 % for

ACCEPTED MANUSCRIPT
P3HT:PCBM-C61 (1:3 wt. %) is reported, Fig 15S [27]. Table 5 summarizes all of the
determined photovoltaic parameters.
As active materials in molecular cables cells both oligo(fulleropyrrolidines) C60-3PEN3
and C60-3PEFe gave very low power conversion efficiencies (order of 10-7), which
increased by two orders of magnitude (to 10-5) by adding an interfacial cathode layer of
PFN (< 5 nm), meaning that there is a better electron recollection to the cathode, but still
with low values.
Table 5. Resume of several tested active layer compositions and the determined
photovoltaic parameters for each material combination.
Jsc
(mA/cm2)
6.37
3.13 X 10-5
2.5 X 10-5
1.0 X 10-4
2.0 X 10-4
0.002
Voc
(mV)
400
FF
Solar cell
composition
η
(%)
1.77
0.69
0.001
0.078
0.260
0.260
0.210
P3HT:PCBM-C60 (1:1)
48
4.3 X 10-7
6.8 X 10-7
1.7 X 10-5
2.5 X 10-5
2.3 X 10-3
C60-3PEN3
C60-3PEFe
130
560
428
604
C60-3PEN3/PFN
C60-3PEFe/PFN
P3HT: C60-3PEFe (1:2)
According to the energy diagram, the ∆LUMOP3HT-C60-3PEFe = 0.32 eV suggests that
charge transfer to P3HT is possible, so, after having evaluated different processing
conditions, which are reported in the legend of Figure 9, the best obtained efficiency for
the heterojunction configuration was of ηAM1.5 = 0.002 %; i.e. a 4 orders of magnitude
higher than that as molecular cable. The Voc has a good value (604 mV), while the
current density 0.023 mA/cm2 is still poor and in general, the photovoltaic values are
lower compared with that of P3HT:PC61BM cell. In this respect, we can identify at least
two factors that limit the intrinsic efficiency of the C60-3PEFe based devices in both
architectures. First, considering Figure 9b, its absorption spectrum in film of c.a. 133 nm
presents the HOMO-LUMO electronic transition at 364 nm that is out of the maximum of
the solar irradiance spectrum, then a drastic shut down of the absorption coefficient (εt)
is observed after 364 nm, meaning low solar light absorption of the film in the visible
range. This is more evident when the photonic recollection (Ih) spectrum of the C60-

ACCEPTED MANUSCRIPT
3PEFe is compared with the solar irradiance. The maxima photonic recollection Ih
~4.53x106Wm-2nm-1 is obtained at 400 nm and afterwards, it dramatically decreases to
the point that at 500 nm the Ih is less than half. The film absorbs 78.4 Wm-2 of the 900
Wm-2 of solar irradiance energy (at AM1.5), representing only the 8.7 % of absorbed
photons. Secondly, low fill factors (FF) are found: 0.078 for C60-3PEFe and 0.210 for the
P3HT:C60-3PEFe devices, Table 5, compared to 0.69 for the P3HT:PCBM device, Fig
15S, ESI. This is likely a consequence of both recombination of charge carriers, in
agreement with the photophysical results and to the poorly ordered morphology of the
active layers. AFM analysis reveals in fact that P3HT:C60-3PEFe device morphology is
composed of agglomerates of different size coexisting with a more planar morphology,
but an amplification of one of these zones reveals deep valleys of c.a. 500 nm in
diameter and of c.a. 25 nm of Z amplitude, Figure 9c-d, evidencing the morphological
discontinuity in the film. In contrast, the AFM image of the P3HT:C60-3PEFe device, Fig
15 Sc, shows a more homogeneous film.
Figure 9. (a) Structure of the P3HT:C60-3PEFe (1:1 w %) solar cell, with thicknesses of
40 nm for the PEDOT:PSS layer and of 130 nm for the active layer. Both layers
underwent annealing at 80 °C for 30 min after their individual deposition, Fields Metal as

ACCEPTED MANUSCRIPT
top contact. (b) Spectra of: solar irradiance (red), photonic recollection (blue) and
absorption coefficient in thin film of C60-3PEFe. (c) Tapping AFM mode image of the
ITO/PEDOT:PSS/P3HT:C60-3PEFe film and (d) is an amplification zone.
The aggregates and the inhomogeneity of the film are responsible for the exciton
diffusion-dissociation in the films not being carried out and that efficiencies of device
constructed with both fulleropyrrolidines are extremely low. In addition, it is likely that the
molecules do not have an adequate balance of electron donor-acceptor character, as the
benzoate is a poor electron acceptor group. This hypothesis could be corroborated when
we compared our results with those found by Nierengarten et al. [28] on fullerene-
oligo(phenyleneethynylene)s bearing bis(octyloxy) as side chains (which are electron
donor groups). The authors report efficiencies in the order of η= x10-2 % for devices in
configuration ITO/PEDOT:PSS/Oligo(PPE-fulleropyrrolidine)/Al. In general, we conclude
that these particular series of oligo(benzoateethynylene)s-either ferrocene or triazene
terminated macromolecules are not adequate for photovoltaic devices. and however are
very thermal stable up to 150 °C, Fig 16S.
3. Conclusions
We succeeded in the synthesis of a novel series of oligomers as precursors and
models of the oligo(fulleropyrrolidines) with increasing length, and, connecting the
ferrocene or triazene as electron-donor group and C60-pyrrolidine as electron-accepting
group in a configuration D→π→A; the benzoates that serve as solubilizing side chains,
have a weak electron-accepting character. All of the ferrocene compounds with and
without fulleropyrrolidines, in solution, oxidize time dependently, becoming insoluble
likely due to the fact that ferrocene not only is a good electron donor but an excellent
light trapping group; phenomenon that was corroborated by cyclic voltammetry, where
either in the oxidation or in the reduction cycle an insoluble precipitate was formed. The
oligomers and oligo(fulleropyrrolidines) bearing ferrocene revealed an electrochemically
reversible one-electron oxidation wave, which shifted anodically with respect to free
ferrocene, but cathodically depending on the length of the π-bridge and of the strength
of the electron-accepting group in the order C60-pyrrolidines > benzoates. In contrast, in

ACCEPTED MANUSCRIPT
the reduction cycle, only one wide wave was observed which shifted anodicaly as a
function of both the length of the π-bridge and of the strength of the electron-accepting
group in the order C60-pyrrolidines < benzoates. These results suggest the existence of
electron transfer between the iron centre of the ferrocene to the electron-accepting
groups through the π-bridge. The photophysical study showed a strong fluorescence
quenching and higher non-radiative constants of the benzoatheethynylene moiety after
functionalization with triazene and even stronger, in the case of ferrocene. The energy
considerations suggest that intramolecular photoinduced charge transfer is possible,
upon direct excitation of the fullerene moiety or indirectly, after energy transfer from the
BzE unit to C60-pyrrolidines, in case of excitation of the conjugated moiety.
Nevertheless, as the absorption range of the oligomers is out of the maximum solar
irradiance and due to inhomogeneous organization in films, photovoltaic devices
present low efficiencies.
4. Experimental Section
4.1 Instruments. 1H (300 MHz), DQF-COSY, DEPT-135, APT, 13C (75.4 MHz) NMR
spectra were obtained at room temperature with a Jeol Eclipse spectrometer using
CDCl3 as solvent and internal reference. The photophysical characterization was carried
out in Aldrich spectroscopic grade CH2Cl2. UV-Vis and fluorescence spectra were
recorded on an Agilent 8453 spectrophotometer and a Perkin Elmer LS50B
spectrofluorimeter, respectively. Fluorescence quantum yield (φ) was obtained
according to literature method [29] with quinine sulphate (0.1M in H2SO4) as standard.
Three different solutions with optical density at the excitation wavelength < 0.1 were
analyzed at 25.0±0.3 °C by using a circulating water bath and the values were
averaged. The excitation wavelength was 10 nm below the absorption maximum.
Lifetimes were obtained by TCSPC (time-correlated single photon counting) with a
Tempro Horiba instrument with a 285 nm nanoLED. A 0.01% suspension of Ludox
AS40 (Aldrich) in ultrapure water was used for the prompt signal. Calibration of the
equipment was realized with a POPOP [1,4-Bis(4-methyl-5-fenil-2-oxazolyl)benzene]
methanol solution (optical density < 0.1 and lifetime of 0.93ns [30]. Data were fit using
DAS6 software available with the equipment. The electrochemical properties of all of the
compounds were investigated via cyclic voltammetry in a Stand cell from Basi, coupled

ACCEPTED MANUSCRIPT
to an ACM Gill AC potentiostat/galvanostat. The system consisted of a conventional
three-electrode cell: glassy carbon as a working electrode (polished with alumina and
diamond powder after each run), Pt wire as the counter electrode, Ag/AgCl as reference
electrode (viability of -35 ± 20 mV against the Calomel electrode). Voltammetry
measurements were performed at room temperature in CH2Cl2 containing Bu4NPF6
(0.1M) as the supporting electrolyte. Prior to recording the voltammograms, all solutions
(~0.5 mmol) were deoxygenated by bubbling nitrogen for at least 15 min. The
experiments were carried out under a nitrogen atmosphere at a scanning rate of 50-100
mV/s for a complete cycle between -3.0 and +3.0 V and in separated windows. The
molecular orbital energies HOMO and LUMO were calculated according to the empirical
relationship [27]: EHOMO[LUMO]= -exp(E(Ox[red]vs. Ag/AgCl))-4.80. Solar cells were fabricated
with the configuration: ITO/PEDOT:PSS/active layer/Al or Field´s metal. The ITO slides
(15-30 ohms) were cleaned in a Branson ultrasonic bath with different solvents: 1)
methylene chloride (10 min); 2) hexane (10 min) and 3) methanol (two cycles of 20 and
40 min, respectively). PEDOT:PSS and the active layers were deposited by spin coating
with a WS-400-6NPP-LITE spin coater from Laurel Technologies. Commercial
PEDOT:PSS suspension was first filtered in a syringe 0.45 micro filter and then spin
coated on the ITO slides at 5000 rpm for 25 s and at 2500 rpm for 1min. The
PEDOT:PSS layer (approximately 20 nm) was dried in an oven at 80 °C for 30 min
before the following deposition step. Active layers consisted of: i) P3HT:PCBM-C60 as
comparative term. In this case, a 1 wt/wt chlorobenzene solution of P3HT:PCBM-C60
was prepared and kept 12h at 60 °C under magnetic stirring. The mixture was
maintained 1h at room temperature under magnetic stirring and then spun at 800 rpm
on the PEDOT:PSS layer to obtain a layer of approximately 100 nm. ii) As molecular
cable, C60-3PEFe was stirred in chlorobenzene and for 12h at 50°C. The mixture was
maintained 1h at room temperature under magnetic stirring and then spun at 1500 rpm
on the PEDOT:PSS layer to obtain a layer of approximately 100 nm. iii) P3HT:C60-
3PEFe: 1 wt/wt chlorobenzene solution of P3HT:C60-3PEFe was dissolved in
chlorobenzene and kept at 50 °C under stirring. The mixture was maintained 1h at room
temperature under magnetic stirring and then spun at 1200 rpm to obtain a layer of c.a.
130 nm. Then, the active layer was annealed at 150 °C for 30 min. Al or Field’s metal

ACCEPTED MANUSCRIPT
(FM) as cathode were deposited by vacuum evaporation (Al) or at 85 °C by using a
heating plate (FM). The I-V curves were obtained with a Keithley 2420 source meter by
illuminating the devices from the ITO side with a 100 mW/cm2 white light from a Solar
Light Co. Model XPS 400 solar simulator with a Xenon lamp and AM1.5 filter. All cells
were prepared and measured under ambient conditions. The film thickness was
determined using a perfilometer Veeco Dektak 6M Stylus with a 12.5 µm radius tip at a
speed rate of 50 µm/s and an applied force of 2 milligrams. Morphological
characterization of the active layers by AFM was performed on a Nanoscope III
DimensionTM 3100 from Digital Instruments with a Pt-coated Si tip (15 nm nominal
radius, model OSCM-PT from Bruker) in tapping mode at a scanning rate (256 lines) of
0.2 Hz.
4.2. Synthesis. Experimental procedures, chemical and physicochemical
characterization of each compound are given in the electronic supplementary
information section.
4.3. Theoretical calculations. Theoretical calculations were carried out using the
Gaussian 09 software using B3LYP functional with the 6-311+G** basis set for C,N,H,O
and m6-31G* one for Fe. The latter provides a more appropriate description of the first-
row transition metals.
Acknowledgments
We wish to acknowledge the Mexican National Council for Science and Technology for
the financial support through the projects CB-2015: 256716 and 256709, National
Laboratory of Graphene. Authors also thank to Guadalupe Méndez and Alfonso
Mercado for their technical help.
Electronic Supplementary Information (ESI) available: experimental
1
13
procedures, H, C, DQF COSY, HETCOR NMR spectra, excitation and fluorescence
spectra, fluorescence time decay profiles of selected oligomers and a P3HT:PC60BM
solar cell device are found at htpp//...

ACCEPTED MANUSCRIPT
References
[1] N. S. Sariciftci, L. Smilowitz, A. J. Heeger, F. Wudl, Science 258 (1992) 1474-1476.
[2] M. Li, K. Gao, X. Wan, Q. Zhang, B. Kan, R. Xia, F. Liu, X. Yang, H. Feng, W. Ni, Y.
Wang, J. Peng, H. Zhang, Z. Liang, H.-L. Yip, X. Peng, Y. Cao, Y. Chen, Nat.
Photonics 11 (2017) 85–90.
[3] W.-Y. Wong, X.-Z. Wang, Z. He, A.B. Djurišić, C.-T. Yip, K.-Y. Cheung, H. Wang,
C.S.-K. Mak, W.-K. Chan, Nat. Mater. 6 (2007) 521.
[4] A.Z. Weller, Phys. Chem. Neue Folge 133 (1982) 93-98.
[5] A. Manzano-Ramírez, E.J. López-Naranjo, W. Soboyejo, Y. Meas-Vong, B. J.
Vilquin, J. Nanomater. 2015 (2015) 1-15.
[6] S. Cataldo, P. Salice, E. Menna, B. Pignataro, Energy Environ. Sci. 5 (2012) 5919-
5940
[7] C. Li, H. Wonneberger, Adv. Mater. 24 (2012) 613–636.
[8] T. Liu, A. Troisi, Adv. Mater. 25 (2013) 1038–1041.
[9] (a) 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 104 (2000) 10174-10190.
(b) R. Gómez, J.L. Segura, Tetrahedron 65 (2009) 540-546.
(c) J.-F. Eckert, J.-F. Nicoud, J.-F. Nierengarten, S.-G. Liu, L. Echegoyen, F.
Barigelletti, N. Armaroli, L. Ouali, V. Krasnikov, G. Hadziioannou, J. Am. Chem. Soc.
122 (2000) 7467-7479.
(d) D.vZhu, Y. Li, S. Wang, Z. Shi, C. Du, S. Xiao, H. Fang, Y. Zhou, Synth. Met.
133-134 (2003) 679-683.
[10] (a) A. Marcos Ramos, S.C.J. Meskers, P.A. Van Hal, J. Knol, J.C. Hummelen,
R.A.J. Janssen, J. Phys. Chem. A 107 (2003) 9269-9283.
(b) J-G. Domínguez-Chávez, E. Cruz-Chávez, I. Moggio, E. Arias-Marín, T. Klimova,
I. Lijanova, M. Martínez-García, Fullerenes, Nanotubes, and Carbon Nanostructures
20 (2012) 249–265.
[11] W.-Y. Wong, C.-L. Ho, Acc. Chem. Res. 43 (2010) 1246-1256.
[12] (a) J.L. Segura, R. Gómez, N. Martín, C. Luo, A. Swartz, D.M. Guldi, Chem.
Commun. (2001) 707–708.

ACCEPTED MANUSCRIPT
(b) D.R. Guldi, A. Swartz, C. Luo, R.G ómez, J.L. Segura, N. Martín, J. Am. Chem.
Soc. 124 (2002) 10875-10886.
(c) T. Gu, P. Ceroni, G. Marconi, N. Armaroli, J-F. Nierengarten, J. Org. Chem. 66
(2001) 6432-6439.
(d) Accorsi, G.; Armaroli, N.; Eckert, J.-F.; Nierengarten, J.-F. Tetrahedron Lett.
2002, 43, 65–68.
[13] (a) A. Lembo, P. Tagliatesta, D.M. Guldi, M. Wielopolski, M. Nuccetelli, J. Phys.
Chem. A 113 (2009) 1779–1793.
(b) M. Wielopolski, C. Atienza, T. Clark, D.M. Guldi, N. Martín, Chem. Eur. J. 14
(2008) 6379–6390.
(c) N. Armaroli, G. Accorsi, Y. Rio, P. Ceroni, V. Vicinelli, R. Welter, T. Gu, M.
Saddik, M. Hollerd, J-F. Nierengarten, New J. Chem. 28 (2004) 1627–1637.
(d) J.N. Clifford, T. Gu, J-F. Nierengarten, N. Armaroli, Photochem. Photobiol. Sci.
5 (2006) 1165-1172.
[14] M. Maggini, G. Scorrano, M. Prato, J. Am. Chem. Soc. 115 (1993) 9798-9799.
[15] S.-G. Liu, L. Shu, J. Rivera, H. Liu, J. –M. Raimundo, J. Roncali, A. Gorgues, L.
Echegoyen, J. Org. Chem. 64 (1999) 4884-4886.
[16] J. Palomero, J.A. Mata, F. González, E. Peris, New. J. Chem. 26 (2002) 291-297.
[17] J.N.Wilson, P.M. Windscheif, U. Evans, M. L. Myrick, U.H.F. Bunz, Macromolecules
35 (2002) 8681-8683
[18] J. A. Mata, S. Uriel, R. Llusar, E. Peris, Organometallics 19 (2000) 3797-3802.
[19] M. Prato, M. Maggini, Acc. Chem. Res. 31 (1998) 519-526.
[20] S. Leach, M. Vervloet, A. Després, E. Bréheret, J. P. Hare, T. J. Dennis, H. W.
Kroto, R. Taylor, D. R. M. Walton, Chem. Phys. 160 (1992) 451-466.
[21] (a) R. Seshadri, C. N. R. Rao, H. Pal, T. Mukherjee, J. P. Mittal, Chem. Phys. Lett.
205 (1993) 395–398.
(b) M. C. Rath, H. Pal, T. Mukherjee, J. Phys. Chem. A 103 (1999) 4993–5002.
[22] (a) H. Zhao, Y. Zhu, C. Chen, J. Zheng, Carbon 50 (2012), 4894-4902.
(b) P.K. Poddutoori, A.S.D. Sandanayaka, T. Hasobe, O. Ito; A. Van der Est, J.
Phys. Chem. B 114 (2010) 14348-14357.
[23] H. Kanato, K. Takimiya, T. Otsubo, Y. Aso, T. Nakamura, Y. Araki, O. Ito, J. Org.