Synthesis and properties of new benzothiadiazole-based push-pull dyes for p-type dye sensitized solar cells

Article abstract of DOI:10.1016/j.dyepig.2017.08.055

We report the synthesis of three new push-pull dyes specially designed for the sensitization of mesoporous p type semi-conductor NiO for the construction of p type dye sensitized solar cells (pDSC). Their structure stems from the conjugation of the well-k

Full text of DOI:10.1016/j.dyepig.2017.08.055

Accepted Manuscript
Synthesis and properties of new benzothiadiazole-based push-pull dyes for p-type
dye sensitized solar cells
Yoann Farré, Mahfoudh Raissi, Arnaud Fihey, Yann Pellegrin, Errol Blart, Denis
Jacquemin, Fabrice Odobel
PII:
S0143-7208(17)31448-1
10.1016/j.dyepig.2017.08.055
DYPI 6217
DOI:
Reference:
To appear in: Dyes and Pigments
Received Date: 6 July 2017
Revised Date: 29 August 2017
Accepted Date: 29 August 2017
Please cite this article as: Farré Y, Raissi M, Fihey A, Pellegrin Y, Blart E, Jacquemin D, Odobel F,
Synthesis and properties of new benzothiadiazole-based push-pull dyes for p-type dye sensitized solar
cells, Dyes and Pigments (2017), doi: 10.1016/j.dyepig.2017.08.055.
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

ACCEPTED MANUSCRIPT
Synthesis and properties of new benzothiadiazole-
based push-pull dyes for p-type dye sensitized
solar cells
a†
a†
a,c
a
a
Yoann Farré, Mahfoudh Raissi, Arnaud Fihey, Yann Pellegrin, Errol Blart, Denis
a,b
a
Jacquemin,* Fabrice Odobel*
a
Université LUNAM, Université de Nantes, CNRS, Chimie et Interdisciplinarité: Synthèse,
Analyse, Modélisation (CEISAM), UMR 6230,2 rue de la Houssinière, 44322Nantes cedex 3,
France. E-mail: Fabrice.Odobel@univ-nantes.fr
b
Institut Universitaire de France, 103 blvd St Michel, 75005 Paris Cedex 5, France. E-mail:
Denis.Jacquemin@univ-nantes.fr
c
Institut des Sciences Chimiques de Rennes, UMR 6226 CNRS, Université de Rennes1, 263
Av. du Général Leclerc, 35042, Cedex Rennes, France.
Abstract
We report the synthesis of three new push-pull dyes specially designed for the sensitization of
mesoporous p type semi-conductor NiO for the construction of p type dye sensitized solar
cells (pDSC). Their structure stems from the conjugation of the well-known tris(carboxy-
arlyamine) anchoring moiety to a benzothiadiazole acceptor through a fluorene (PP1) or
thiophene-fluorene linker (PP2). Appending a NDI secondary acceptor onto PP2 yielded a
dyad (PP2-NDI). The push-pull structure is here particularly adapted to the sensitization of p-
1

ACCEPTED MANUSCRIPT
type semi-conducting materials. All three dyes were investigated by spectroscopic and
electrochemical techniques, and their electronic properties were modelled by TD-DFT. In all
cases, the new sensitizers exhibit adequate absorption features (spanning from 400 nm to 600
-1
-1
nm, extinction coefficients between 20000 and 27000 M .cm ) and their thermodynamic
parameters are all in favor of efficient interfacial charge transfers involving NiO. pDSC were
constructed with all three dyes; high photoconversion efficiencies (PCE) were reached,
2
especially in the case of dyad PP2-NDI (Jsc = 5.90 mA/cm , Voc = 143 mV, FF = 33.8%,
PCE = 0.29%), due to a combination of outstanding electronic properties and increased
charge separation state lifetime. Tandem DSC (tDSC) composed of a TiO photoanode
2
sensitized by the dye D35 and a NiO photocathode sensitized by PP2-NDI were assembled
-
-
-
with two different electrolytes (I /I or T /T based electrolytes). The latter gave PCE among
3
2
2
the best ever reported for tDSC (Jsc = 5.90 mA/cm ; Voc = 896 mV; FF = 63%, PCE = 2.80%
-
-
in the case of the I /I electrolyte, Jsc = 4.52 mA/cm², Voc = 840 mV, FF = 60% and PCE =
3
-
2
.27% for T /T based devices) unraveling the significance of the new dyes presented in this
2
study.
Keywords: push pull dyes; benzothiadiazole; fluorene linker; dyad; dye sensitized solar cells;
tandem cells
1
. Introduction
The conversion of solar light into a usable form of energy is one of the greatest challenges
1
ever tackled by the scientific community. Renewable, free and rather well distributed over
the globe, sunlight, if mastered, would doubtlessly put an end to the hegemony of fossil fuels
and the related energy crisis. Photovoltaic devices are certainly the most advanced artificial
2
systems able to directly convert light into electricity. Among them, dye sensitized solar cells
(DSC) are particularly appealing devices owing to their stability, high efficiency, their
aesthetic features and their short payback time, much smaller than the commercialized silicon
3
based devices. The latter are even inferior to DSC in many regards, such as flexibility,
4
transparency and increased efficiency in low light conditions. Classical DSC are based on the
sensitization of a wide band gap n type semi-conductor, usually mesoporous TiO , by a dye
2
3
,5,6
exhibiting adequate electronic properties.
Upon light soaking, the dye injects electrons in
2

ACCEPTED MANUSCRIPT
the conduction band of TiO while a redox mediator regenerates the photo-oxidized
2
7
chromophore. Since they were brought to center stage by Grätzel and O’Regan in 1991, an
impressive progression of the photoconversion efficiencies (PCE) has been experienced, from
8
7
% to ca. 14% today. Since year 2000, a new kind of DSC has emerged based on the
9
-12
sensitization of a mesoporous p-type semi-conductor (such as NiO).
Those p-type dye
sensitized solar cells (pDSC) are based on the injection of holes from a photo-excited dye into
the valence band of the semi-conducting material. Importantly, pDSC and nDSC can be
associated together to design a tandem dye sensitized solar cell (tDSC) featuring theoretical
13,14
efficiencies above the Schockley-Queisser limit.
Previous works have shown in particular
1
2,15-19
that tDSC exhibit larger photovoltages, independent of the redox mediator.
High
photocurrents are attainable too if care is taken to use different dyes for each semi-conductor,
exhibiting complementary absorption spectra in order to avoid screening of light by one dye
1
3,18
or the other.
The mandatory step for constructing an efficient tDSC is that parent nDSC
and pDSC are both efficient and very importantly that both short circuit currents are matched.
While impressive photoconversion efficiencies (PCE) have been obtained for nDSC, pDSC
still lag behind because of the built-in weaknesses of NiO (color, low conductivity) and
20-23
counterproductive phenomena such as geminate and interfacial charge recombinations.
Significant work today is devoted to improving the PCE of pDSC, by fishing out a better
2
4-31
17,25,32-36
semi-conducting material than NiO,
determining new efficient redox mediators,
The latter obviously play a major part in the overall
mechanism of a working pDSC. Within this realm, push pull dyes have early shown their
or
3
6-50
synthesizing new efficient dyes.
4
4,45,51-57
qualities as sensitizers for pDSC.
They exhibit a strongly allowed donor to acceptor
transition affording huge extinction coefficients; besides their color can be tuned by
modifying the nature of the donor and acceptor groups. Additionally, if the anchor is tethered
to the donor side, the hole injection is properly vectorialized, i.e. the very structure of a push
pull sensitizer supports hole injection. The localization of the HOMO close to the semi-
conductor’s surface and of the LUMO as far as possible from the latter is a real advantage of
push-pull sensitizers to favor hole injection and retard geminate charge
1
4,44,50,52,56,58
44
recombination.
Push pull dyes are epitomized by the P1 dye (chart 1), the only
dye for pDSC which is commercially available. Other push pull dyes have since been isolated,
1
8,46,47,59
some displaying greater PCE than P1.
3

ACCEPTED MANUSCRIPT
Herein we present our own contribution in this field, presenting two new push pull dyes PP1
6
0,61
and PP2 (chart 1) based on the association of the well-known trisarylamine anchor (TPA)
5
3,62,63
and the benzothiadiazole acceptor (BTD).
TPA has been widely used as anchoring
group in many dyes for pDSC because it favors hole injection thanks to its electron-rich
character. Besides, the presence of two COOH anchors provides additional stability to the dye
5
5,58
monolayer.
Regarding the acceptor fragment, BTD shows excellent properties that
6
2-65
materialized into a collection of successful opto-electronic devices,
including sensitizers
53
for pDSC. In particular BTD exhibits a reversible reduction potential of ca. -1.2 V vs. SCE,
providing sufficient driving force for regeneration of the photo-reduced dye by a great many
redox mediators. In the case of PP1, the TPA donor is linked to the BTD acceptor by a
fluorene spacer which is known to favor the electronic communication between the two
6
6,67
ends.
Moreover, the presence of 2-ethylhexyl chains on the fluorene moiety provides
solubility to the dyes and could passivate the surface of the semi-conductor by preventing the
redox mediator from approaching too close and curbing therefore the interfacial charge
recombination. In the case of PP2, a thiophene moiety is inserted between the BTD and
fluorene fragments; increasing the distance and conjugation between donor and acceptor sides
entails a rise of the dipole and thus an increase of the extinction coefficient while shifting the
transition to the red, both appreciable features for a performant sensitizer. Moreover, the
probability of geminate charge recombination decreases as the acceptor (where the electron
4
0,50
density is localized after hole injection) is farther away from the semi-conductor’s surface.
Chart 1. Structures of P1 and of the dyes investigated in this work.
4

ACCEPTED MANUSCRIPT
In order to further curb down the extent of geminate charge recombination, a secondary
electron acceptor (naphthalene diimide, NDI) was affixed to PP2 to yield PP2-NDI. Within
such a dyad, a secondary “dark” electron transfer from the photo-reduced chromophore onto
the secondary acceptor shifts the electron farther away from the semi-conducting material
3
7,38,68,69
resulting in longer lived charge separation states by several orders of magnitude.
The syntheses, electronic properties of those three new dyes are disclosed in this contribution,
as well as the photovoltaic performances of associated p-type photovoltaic devices. We found
that the push-pull nature of the dyes entailed the presence of an intense charge transfer
transition in the visible domain, assisting besides the hole injection in the valence band of
NiO; appending a secondary acceptor to the best performing push-pull dye afforded the
highest PCE in the series because of an improved photo-induced interfacial charge separated
state lifetime.
2
. Results and discussion
.1. Synthesis of the dyes
2
The synthesis of the three new sensitizers PP1, PP2 and PP2-NDI stems from building block
and a stannylated compound derivatized from benzothiadiazole (6 or 9) (Schemes 1-2).
3
PP2-NDI was synthesized from compound 11, which is an intermediate of the synthesis of
PP2 (Scheme 3). Compound 3 was obtained from the commercially available building blocks
1
and 2 (Scheme 1), via a Buchwald-Hartwig cross-coupling reaction using Pd(dba) as
2
precatalyst and XPhos as ligand. The fluorene derivative 1 was introduced in excess to
promote the monocoupling reaction and the desired compound 3 was isolated with a yield of
3
2% (based on the limiting amine reagent 2) (Scheme 1). Despite the excess of fluorene
derivative, the biscoupling product was isolated with a yield of 58%.
Scheme 1. Synthesis of compound 3. Reagents and conditions: a) Pd(dba) , XPhos, t-BuOK,
2
toluene, 100°C, 20h, 32%.
5

ACCEPTED MANUSCRIPT
The two stannylated reactants 6 and 9 were obtained from the reaction between the
brominated derivatives 5 and 8, and hexamethyldistannane by a similar catalytic way using
Pd(PPh ) as catalyst in DME or toluene, respectively (Scheme 2). It should be noted that
3
4
compound 9 was purified on triethylamine-neutralized silica gel to avoid hydrolysis of the
stannane.
Scheme 2. Synthesis of compounds 8 and 9. Reagents and conditions: a) Pd(PPh ) Cl , THF,
3
2
2
8
0°C, 24h, 20% for compound 5, 80% for compound 7 b) Pd(PPh ) , Sn Me , DME, 85°C,
3 4 2 6
2
4h, 66% c) NBS, CH Cl , RT, 16h, 32% d) Pd(PPh ) , Sn Me , toluene, 100°C, 5h, 86%.
2
2
3 4
2
6
A Stille coupling between building block 3 and stannylated derivatives 6 or 9 afforded
respectively compound 10 and 11 in about 50% yield (Scheme 3). For compound 11, the
classical Stille conditions (Pd(PPh ) in toluene at 100°C) gave a satisfactory yield. However,
3
4
they were not well adapted to synthesize compound 10, probably due to the steric hindrance
generated by the benzothiadiazole moiety on the ortho position of the stannylated compound
6
. Mee et al. have studied the reactivity for Stille reaction of some less reactive
7
0
arylstannanes, and they discovered a synergic catalytic system using CuI in combination
with CsF, affording good results in difficult cases. We have used these conditions to realize
the coupling between 3 and 6 yielding compound 10 with a yield of 53%.
6

ACCEPTED MANUSCRIPT
Scheme 3. Synthesis of dyes PP1 and PP2. Reagents and conditions: a) PdCl , P(t-Bu) , CuI,
2
3
CsF, DMF, 50°C, 18h, 53% b) TFA, CH Cl , RT, 4-5h, 99% c) Pd(PPh ) , toluene, 110°C,
2
2
3 4
2
4h, 69%.
The hydrolysis of the tertbutanol ester groups was realized by action of TFA in
dicholoromethane and allowed isolating the sensitizers PP1 and PP2 in high yields as yellow
and red powders, respectively.
The synthesis of PP2-NDI was accomplished in three steps from the intermediate compound
1
1. The first step is a selective iodination reaction, already used for other thiophene
7
1,72
derivatives
and afforded 12 in 84% of yield. Sonogashira cross-coupling reaction between
1
2 and ethynyl NDI 13 afforded the dyad PP2-NDI after hydrolysis of the tertbutanol ester
groups of 14 with TFA.
7

ACCEPTED MANUSCRIPT
Scheme 4. Synthesis of dyes PP2-NDI. Reagents and conditions: a) Hg(OCp)2,
CHCl /AcOH, 1h, RT and then I , 3h, RT, 84% over two steps; b) TFA, CH Cl , RT, 4-5h,
3
2
2
2
9
9%; c) Pd(PPh ) , toluene, 110°C, 24h, 69%. c) TFA, CH Cl , RT, 5h, 96%.
3 4 2 2
2
.2. Electronic UV-visible absorption and emission spectroscopy
The absorption and emission spectra of all dyes were recorded in dichloromethane solution at
room temperature (Figure 1). The optical data including wavelengths of maximal absorption
(
λ ) with extinction coefficient (ε), wavelength of maximal emission (λ ) and 0-0 energy
abs em
level of the lowest singlet excited state (E ), are collected in Table 1.
0
0
The UV-Visible absorption spectra of PP1 and PP2 in dichloromethane exhibit an intense
absorption band in the visible, respectively at 438 and 485 nm attributed to a π-π* transition
with a moderate charge transfer character (see calculations below) with molar extinction
-1
-1
coefficient (ε) values of ca. 20 000 and 27 000 M .cm at λmax. It should be noted that the
extension of the π-conjugated system with a thiophene unit (from PP1 to PP2) induces a large
bathochromic shift of around 50 nm and an increase of the molar extinction coefficient. These
properties enable enhancing the ability of PP2 to harvest sunlight compared with PP1. Both
dyes show another absorption band around 350-380 nm with a molar extinction coefficient of
8

ACCEPTED MANUSCRIPT
-1
-1
around 50000 M .cm . The theoretical results show this band is a π-π* transition globally
delocalized on the whole dye, with a limited charge transfer.
Figure 1. Electronic absorption (straight line) and normalized steady state emission (dashed
lines) spectra of the sensitizers PP1, PP2 and PP2-NDI recorded in dichloromethane solution.
PP2 and PP2-NDI exhibit a similar absorption spectra. The main change consists in the
appearance of two narrow bands at 361 and 380 nm on the spectrum of PP2-NDI, which can
be attributed, after analyzing the theoretical spectrum obtained with quantum calculations, to
the contribution of NDI localized absorption features, mixed with the bands in the UV
described above for PP1 and PP2. For this reason, that band reaches a global molar extinction
-1
-1
coefficient of around 80000 M .cm . More importantly, the main absorption band in the
visible is significantly broader and slightly red-shifted compared to PP2. Accordingly, the
NDI unit is electronically decoupled from the rest of the molecule because of a well-known
7
3
orbital node on the nitrogen atom of the diimide group. It is thus unlikely that the NDI be
responsible for the broader and bathochromically shifted absorption band of PP2-NDI vs.
PP2. However, the phenylacetylene spacer formally bridging PP2 to the NDI moiety in the
dyad extends the conjugation of the π system and account for the differences observed
between PP2-NDI and PP2. This red-shift is actually validated by our quantum calculations,
returning a LUMO+1 energy level for PP2-NDI lower than the LUMO of plain PP2 (by 0.13
eV), while their respective HOMO levels remain unchanged.
9

ACCEPTED MANUSCRIPT
PP1 and PP2 are both fluorescent at room temperature in diluted dichloromethane solutions
when excited at λmax and show an emission band with maxima at 596 and 621 nm,
respectively (Figure 1). Unsurprisingly, λem and λabs follow the same trend for PP1 and PP2.
-1
The Stokes shifts are rather large (respectively 6051 and 4516 cm , table 1) indicating a
significant redistribution of the electron density upon excitation and supporting the
assumption that the absorption band in the visible possesses a significant charge transfer
character. The addition of the NDI unit almost completely quenches the fluorescence. By
analogy to other work on related dyads, we could hypothesize that the singlet-excited state is
+
- 37,49,50,74
quenched by photoinduced electron transfer to form PP2-NDI .
Table 1. Wavelengths of maximal absorption (λ ) with extinction coefficient (ε),
abs
wavelength of maximal emission (λ ) recorded at room temperature in dichloromethane,
em
associated Stokes shift and zero-zero energy level of the lowest singlet excited state (E ).
0
0
Stokes shift
-
1
-1
a
Dyes
λabs /nm (ε/M cm )
λem /nm
E00 / eV
2.41
-1
/
cm
PP1
PP2
596
6051
4516
4560
4
4
38 (20000); 352 (46600)
85 (27100); 364 (61200)
621
651
2.23
2.15
PP2-NDI
5
02 (26400) 380 (77500); 361 (81300)
a
calculated according to the equation: E = 1240/ λinter, with λinter being the wavelength at the
00
intersection of the normalized absorption and emission spectra.
2
.3. Electrochemical study and electron transfer driving forces
The three new sensitizers were studied by cyclic voltammetry (Table 2) to determine their
redox potentials and estimate the hole injection (∆G° ) and dye regeneration (∆G° ) driving
inj
reg
forces (Table 2).
For each dye, an oxidation process occurs around 1.0 V vs. SCE with a generally non-
1
6,68
reversible or weakly reversible wave attributed to the arylamine moiety.
The HOMO level
seems to be weakly influenced by the introduction of the thiophene unit between the fluorene
connecting group and the BTD acceptor (0.05 V shift between PP1 and PP2). A similar shift
though of larger amplitude (0.17 eV) was obtained when examining the DFT-computed
1
0

ACCEPTED MANUSCRIPT
HOMO energy levels. PP1 and PP2 are respectively reversibly reduced at -1.4 and -1.3 V vs.
SCE. The comparison of these values indicates the stabilization of the LUMO level of PP2
compared to PP1, in agreement with the red-shifted band of PP2 observed on the absorption
spectra. This is also the case when considering the PP1 and PP2 LUMO theoretical levels
obtained with our quantum scheme, located respectively at -1.80 and -1.96 eV. For the
sensitizer PP2-NDI, the two first reversible reduction processes take place on the NDI moiety
around -0.6 and -1.0 V vs. SCE. Unsurprisingly, the DFT LUMO localized on the NDI (see
below) is also shifted to lower energy (-2.63 eV) compared to its PP1 and PP2 counterparts.
Using both electrochemical and spectroscopic data, we estimated the Gibbs free energies for
the various charge transfer processes. It can be noted that the hole injection process into NiO
valence band for each sensitizer display sufficient exergonicity (superior to 500 meV, Table
2
). Besides, PP1 and PP2 feature very large regeneration driving forces with common redox
mediators (ca. 1 eV for iodine based electrolytes, more than 1.5 eV for disulfide/sulfide based
redox mediators). When appending NDI to PP2, a severe drop of the photo-reduced dye
regeneration is naturally experienced, given that hole injection should be followed by a quick
electron shift from the PP2 moiety to the NDI unit. The loss in regeneration driving force is
usually counter-balanced by an important rise of the interfacial charge separation state
lifetime, a beneficial feature when the regeneration process is sluggish. As reported in
previous work, geminate charge recombination (i.e. between holes in NiO and the photo-
reduced sensitizer) is indeed a very deleterious process (among others), which prevents pDSC
38,75-77
reaching high output performances comparable to conventional nDSC.
Grafting
secondary acceptors on efficient dyes proved to be a very powerful method to decrease
geminate charge recombination, at the expense though of the regeneration driving
3
7,49,50,74
force.
Given the estimation of the driving forces for the various charge transfer processes dyes PP1,
PP2 and PP2-NDI should be efficient sensitizers for pDSC.
1
1

ACCEPTED MANUSCRIPT
Table 2. Redox potentials recorded by cyclic voltammetry at room temperature in
dichloromethane/DMF [95:5] solution with Bu NPF (0.1 M) as supported electrolyte and
4
6
referenced versus saturated calomel electrode (SCE). Calculated Gibbs free energies for hole
injection (∆G° ) and dye regeneration (∆G°reg
)
inj
E1/2
b
c
d
+
-
-
a
-
∆G°inj
∆G°reg
∆G°reg
E (S /S) (V)
Ε (mV)
E1/2 (S/S ) (V)
(NDI/NDI )
(V)
E (S*/S )
(V)
-
-
Dyes
with I
3
with T
2
/T
∆
∆Ε (mV)
(eV)
(eV)
(eV)
∆
Ε (mV)
PP1
PP2
1.01 (90)
0.95 (70)
-1.40 (110)
-1.29 (80)
-
1.01
0.94
-0.71
-0.64
-1.08
-0.97
-0.27
-1.65
-1.54
-0.88
-
e
PP2-NDI
1.04
-1.20 (100) -0.59 (90)
0.95
-0.65
a
-
-
b
Calculated according to the equation: E (S*/S ) = E (S/S ) + E . Calculated according to
1
/2
1/2
00
-
c
the equation: ∆G° =E (NiO)- E (S*/S ) with E (NiO) = 0.30 V vs. SCE. Calculated
inj
BV
Red
BV
-
-
-●
-
-●
according to the equation: ∆G = E (S/S ) - E(I /I ) with E(I /I ) = - 0.32 V vs. SCE.
reg
Red
3
2
3
2
d
-
-
-
Calculated according to the equation: ∆G = E (S/S ) - E(T /T ) with E(T /T ) = 0.25 V vs.
reg
Red
2
2
e
SCE. Non reversible process, peak potential.
2
.4. Quantum chemical calculations
TD-DFT calculations were undertaken in order to have a deeper understanding of the optical
and electrochemical properties of dyes PP1, PP2 and PP2-NDI. The key results are gathered
in Table 3. The frontier orbitals of interest are represented in Figure 2. The HOMO and
HOMO-1 are delocalized over the whole push-pull structure and importantly on the anchoring
2
2
triphenylamine moiety, which is favorable for hole injection in the valence band of NiO.
Conversely, the LUMOs are centered on the benthiadiazole acceptor end of the dyes PP1 and
PP2, thus shifting the electron density away from the semi-conductor surface and limiting
geminate charge recombination. In PP2-NDI, the LUMO is purely centered on the NDI
3
8,50,78
moiety as expected for this kind of dyad-dye,
which corroborates the electrochemical
measurements. The calculated absorption spectra reproduce quite well the experimental data,
the computed λabs values following the experimental ranking. The transitions are similar for
the three dyes, mainly HOMO and HOMO-1 to LUMO for PP1 and PP2, and HOMO and
1
2

ACCEPTED MANUSCRIPT
HOMO-1 to LUMO + 1 for PP2-NDI since the LUMO is localized on the NDI in this case;
the latter is completely isolated from the rest of the molecule from the point of view of the
electronic communication and therefore does not participate in the intense transition in the
visible due to a poor overlap with the occupied orbitals. Ground and excited state dipole
moments indicate incidentally that the structures for PP1 and PP2 are adequately designed to
1
4
favor photo-induced hole transfer into NiO. Interestingly, the computed transferred charge
CT
q
upon visible light irradiation is almost constant for the three dyes but the charge transfer
CT
distance d is substantially higher for PP2 than PP1 or PP2-NDI, a promising feature for an
efficient photo-induced hole injection into NiO. To sum up, the computed structures respect
the requested features in the design of efficient dyes for pDSC.
At this point, changes in the hole injection process between the three dyes can be intuited and
directly related to photovoltaic efficiency. In particular the impact of the NDI unit can be
clearly highlighted. First, we note that the ground and excited state dipole moments values
indicate that the structures for PP1 and PP2, and PP2-NDI are adequately designed to favor
1
4
photo-induced hole transfer into NiO, and, that, the highest excited state dipole moment is
CT
obtained for PP2. The computed transferred charge q upon visible light irradiation is
CT
almost constant for the three dyes but the charge transfer distance d is again substantially
higher for PP2 than PP1 or PP2-NDI. Thus PP2 dye is expected to yield the most efficient
hole creation after vertical excitation among the series, and the presence of the NDI unit is not
modifying this "FC" picture. However, when considering the presence of a NDI-localized
LUMO in PP2-NDI (Figure 2), one can expect, as evoked earlier, that the electron will be
transferred to this secondary acceptor unit where it will be trapped, away from the
benthiadiazole (LUMO+1). This yields to an improved separation of the electron/hole pair
and, consequently, an important drop in the charge recombination. Assuming fast electron
transfer, PP2-NDI should therefore outperform both PP1 and PP2. To sum up, all the
computed structures respect the requested features in the design of efficient dyes for pDSC.
Yet, both quantitative and mechanistic discrepancies are highlighted for the hole injection and
charge recombination process within the series, that are further clarified with the photovoltaic
measurements.
1
3

ACCEPTED MANUSCRIPT
Table 3. Computed Wavelength of Maximal Absorption (λabs) with Oscillator Strengths (f),
CT
Ground-state and (first) Excited-state Dipole Moment (µ), Charge-Transfer Distance (d ),
CT
and Amount of Transferred Charge (q )
CT
CT
Dye
λabs / nm (f/au)
422 (0.87)
µ/D
d /Å
q /e
PP1
1.19/4.61
0.30/6.44
0.87/5.01
1.50
2.42
1.73
0.60
0.58
0.57
PP2
474 (1.12)
PP2-NDI
501 (1.67)
Figure 2. a) Frontier molecular orbitals (isovalue: 0.02 au) and b) transition densities of the
first excited-state (isovalue: 0.0004 au) for the different dyes. Green (white) areas correspond
to a gain (loss) of electronic density.
3
. Photovoltaic measurements
1
4

ACCEPTED MANUSCRIPT
As mentioned above, all new dyes presented in this study are able to thermodynamically
inject holes in the valence band of NiO. pDSC were thus prepared in order to evaluate the
performances of PP1, PP2 and PP2-NDI in NiO based hybrid photovoltaic devices, using
three different electrolyte compositions. All details concerning the preparation of the cells are
gathered in the experimental section. Briefly, optimized mesoporous 3.5 µm thick NiO
electrodes were prepared by screen printing a home-made NiO ink on FTO substrates.
Importantly, a dense layer of NiO was carefully deposited on the FTO substrates in order to
7
9
improve the ohmic contact between FTO and NiO. The as-obtained electrodes were sintered
at 400°C for 30 minutes and a last nickel acetate treatment was performed on the NiO
8
0,81
electrodes in order to passivate the superficial defects.
The electrodes were dipped in the
dye solution overnight, then washed and dried. They were finally assembled with a platinum
coated FTO substrate using a 25 µm hotmelt spacer and an electrolyte was introduced by
vacuum back-filling. The performances of the pDSCs recorded under AM1.5 simulated solar
light are gathered in Table 4 for the different electrolytes, and the corresponding J/V curves
are given in figure S1 . Among the three dyes, PP2-NDI undeniably yields much better
photoconversion efficiencies than PP1 and PP2 due to enhanced Voc and particularly high
Jsc. The latter is illustrated by the higher external quantum efficiency (EQE) spectra obtained
for PP2-NDI (Figure 3). Cells prepared with PP1 and PP2 feature similar performances, PP2
being nevertheless substantially better than PP1, consistently with abinitio calculations,
owing to a higher short circuit current density (Jsc, 25% better for PP2 based device in the
case of the iodine based electrolyte). The latter parameter is given by the following formula:
Jsc = ηcoll x η x η x LHE
inj
reg
where η , η , η are respectively the charge collection, injection and regeneration quantum
coll
inj
reg
yields and LHE is the light harvesting efficiency. Given the thermodynamic parameters
involved for PP1 and PP2 (Table 2), one can fairly consider that η , η and η are similar
coll
inj
reg
for both PP1 and PP2 based photocathodes. However, PP2 is endowed with a higher
extinction coefficient than PP1 and red-shifted absorption bands leading to an enhanced light
harvesting efficiency (LHE) for the PP2|NiO photocathode, most likely accounting for the
improved Jsc. The larger charge transfer distance for PP2 (computational chemistry section,
Table 3) is an appreciable factor too to explain the better efficiency of the latter dye than PP1.
As mentioned above, the best dye in the series is by far PP2-NDI, mostly because of an
impressive average Jsc of 5.90 mA/cm² and a large Voc of 143 mV. Although PP2-NDI
1
5

ACCEPTED MANUSCRIPT
displays slightly better optical properties than PP2, those cannot account for the blatant
improvement of the photocurrent, nor could a better ηinj since injection driving forces are
similar for both dyes. Incidentally, dyad PP2-NDI features the weakest regeneration driving
force because of the appended NDI acceptor moiety (Table 2). As intuited from the quantum
calculations, after the photo-induced hole injection, a quick electron shift is expected to occur
from the PP2 fragment to the NDI secondary acceptor as observed before in dyadic
37,49,50,74
sensitizers.
This charge shift allows an increase in the interfacial charge separation
state at the expense of the regeneration driving force, which is here around 700 meV smaller
for PP2-NDI than PP1 or PP2. Given the obtained performances, such a sacrifice was a
(+)
(-)
necessary evil: the increased lifetime of the charge separated state NiO |PP2-NDI is
38,50,78
synonymous with a dramatically curbed geminate charge recombination.
ηcoll is
therefore much improved explaining the higher Jsc, as well as the significantly increased Voc
for PP2-NDI devices compared to PP1 and PP2 ones.
Iodine based electrolytes are not optimal for DSC in general and NiO based pDSC in
-
particular. I is indeed corrosive (limiting the lifetime of the DSC), quite intensely colored
3
-
-
(
partly screening the dyes from solar light), E(I /I ) is too positive limiting the attainable Voc
3
-
-
and the I to I oxidation reaction is catalyzed by NiO (increasing the extent of the interfacial
3
3
2-35
charge recombination). Cobalt and iron complexes proved to be interesting alternatives.
3
+
2+
n+
Within this frame, we tested the Co1 /Co1 based electrolyte (where Co1 is the tris(4,4’-
(n+)
diterbutyl-2,2’-bipyridine)cobalt
complex, cf. Figure S3 for the structures of the
compounds) with the three new dyes. Importantly, this redox mediator is only compatible
(+)
(-)
with a photocathode where the interfacial charge separation state NiO |Dye is long
3
5,37
lived.
Cells elaborated with dyes PP1 and PP2 thus performed poorly with this electrolyte
because of an assumed very fast geminate charge recombination. On the other hand, pDSCs
built with dyad PP2-NDI featured a respectable Jsc of 1.72 mA/cm² (Table 4), in accordance
3
+
with a rather efficient dye regeneration by Co1 . In this case, a remarkable Voc of ca. 340
3
8,50,78
mV was obtained agreeing with previously published work.
The improved output
potential is assigned to the severely decreased interfacial charge recombination between holes
2
+ 35,37,82
in the valence band of NiO and the negative charges borne by the very bulky Co1 .
This result is additional evidence that the interfacial photo-induced charge separated state is
long lived, and therefore that the “dark” electron shift from the PP2 moiety to the NDI unit is
efficient.
1
6

ACCEPTED MANUSCRIPT
-
In 2010, redox mediators based on a tetrazole-thiolate and its disulfide dimer (T /T couple)
2
-
-
83
(
Figure S3) “rivaled” the I /I couple within a classical Grätzel cell. Shortly after that, the
3
-
T /T couple was successfully tested in pDSCs, showing great promise towards improved Voc
2
because of decreased interfacial charge recombination processes and a 70 mV less positive
-
17,25
-
potential E(T /T ).
Additionally, the T /T electrolyte is almost colorless allowing more
2
2
light to be collected by the sensitizer. PP1, PP2 and PP2-NDI were therefore tested with the
latter and the same trends than above were monitored, with PP1 and PP2 lagging well behind
PP2-NDI, likely for the same reasons as those applying to iodine and cobalt based
electrolytes (e. g. fast geminate charge recombination). Interestingly, Voc were raised in all
cases (by 55, 42 and 69 mV for PP1, PP2 and PP2-NDI based pDSC respectively) as
1
7
-
previously observed. This can be assigned to the less positive redox potential of T /T , but a
2
quick examination of the J/V curves recorded in the dark revealed that the dark current
-
- -
substantially decreased too with T /T compared to I /I (Figure S2).
2
3
Table 4. Mean Photoelectrochemical Metrics of the p-DSCs sensitized with PP1, PP2 and
-
-
-
PP2-NDI dyes and employing either the triiodide/iodide (I /I ), disulfide/sulfide (T /T ) or
3
2
3
+
2+
2
cobalt (Co1 /Co1 ) electrolytes recorded under AM1.5 G Simulated Sunlight (1000 W/m ).
Dye
PP1
Electrolyte Jsc (mA/cm²)
Voc (mV)
113.6 ± 3.3
168.9 ± 7.8
115.6 ± 5.4
157.5 ± 0.8
143.1 ± 2.5
212.5 ± 7.8
342.2 ± 11.3
FF (%)
η (%)
-
-
-4
I /I
1.53 ± 0.03
1.60 ± 0.09
2.15 ± 0.05
1.82 ± 0.03
5.90 ± 0.01
4.31 ± 0.19
1.72 ± 0.01
31.9 ± 0.4
30.5 ± 0.7
30.6 ± 0.7
31.5 ± 0.1
33.8 ± 0.3
33.9 ± 0.5
39.7 ± 0.3
0.055 ± 10
3
-
T /T
2
0.082 ± 0.002
-
-
-4
PP2
I /I
0.075 ± 10
3
-
T /T
2
0.090 ±0.001
0.29 ± 0.005
0.23 ± 0.01
0.31 ± 0.01
-
-
PP2-NDI
I /I
3
-
T /T
2
III
II
Co /Co
1
7

ACCEPTED MANUSCRIPT
Figure 3. External quantum efficiencies (EQE) of the dyes in NiO based DSCs.
The interesting performances of PP2-NDI, and in particular the high average Jsc obtained
-
-
with an I /I electrolyte based pDSC prompted us to construct tandem cells incorporating a
3
NiO|PP2-NDI half-cell such as those presented above. A tandem DSC (tDSC) is based on a
photoanode and a photocathode facing each other and bathed in the same electrolyte. The
photoanode was made of on a 4 µm thick TiO mesoporous electrode sensitized by well-
2
known orange D35 push pull dye (cf figure S3). The latter was chosen because of its intense
8
4-
absorption features and its high efficiency and compatibility with various redox mediators.
8
6
Besides D35 is an organic dye like PP2-NDI; this is advantageous compared to heavy
metal based dyes such as N719, from the environmental and financial points of view.
-
- -
First, the performances of the sub-cells prepared with the T /T and I /I electrolytes were
2
3
evaluated under AM 1.5 simulated solar light. Typical J-V curves are given in Figure 4 and all
the results are gathered in Table 5.
2
Table 5. Metrics of the tandem and sub-cells recorded under simulated AM1.5 (100 mW/cm )
-
-
-
with I /I and T /T based electrolytes.
3
2
electrolyte
Jsc (mA/cm²)
6.48
Voc (mV)
175
FF (%)
34
η (%)
0.38
0.31
3.72
2.6
-
-
NiO|PP2-NDI
I /I
3
-
-
-
T2/T
4.50
204
34
-
-
TiO |D35
I /I
7.41
764
65
2
3
T2/T
6.94
641
61
-
-
tDSC
I /I
5.07
896
63
2.80
2.27
3
T2/T
4.52
840
60
1
8

ACCEPTED MANUSCRIPT
The EQE of each sub-cell is given in Figure S4 and confirms that the photoanode is able to
fill the absorption gap (between 420 and 520 nm) left by the PP2-NDI|NiO photocathode. A
substantial overlap is nevertheless observed and explains that the tandem cells display a
slightly smaller Jsc than each sub-cell taken separately (table 5, figures 4 and S5), regardless
of the electrolyte. However, a quite high output potential of almost 900 mV is obtained in the
-
-
case of the I /I electrolyte roughly corresponding to the sum of each sub-cell’s Voc. A
3
-
smaller Voc of 840 mV was measured for T /T based tDSC (figure S5) mainly because of the
2
smaller photopotential generated by TiO |D35 sub-cell as a consequence of higher interfacial
2
charge recombination. Let us lay stress under the fact that D35 based half cells are very
performant although only a thin layer of mesoporous TiO was used; the high Voc obtained
2
are particularly beneficial for the elaboration of a tandem cell, justifying again the use of D35
in our case. Nevertheless, the high Jsc and Voc measured with these tandem cells highlight
their added value, especially when care is taken to use sensitizers showing complementary
absorption spectra and exhibiting high performances in half cells. Overall, an impressive
-
-
photoconversion efficiency of 2.80% was measured for the TiO |D35 / I /I / PP2-NDI|NiO
2
3
1
2,16-19,87
tDSC, one of the highest reported so far for tDSC.
8
6
4
2
0
.5
.5
.5
.5
.5
V (mV)
1000
-100
0
100
200
300
400
500
600
700
800
900
-1.5
Tandem in light
D35-in light
Tandem in dark
PP2-NDI
D35 in dark
PP2-NDI
1
9

ACCEPTED MANUSCRIPT
-
-
Figure 4. J-V characteristic for the TiO |D35 / I /I / PP2-NDI|NiO tDSC (purple lines) and
2
3
associated half cells (TiO |D35 in yellow and PP2-NDI|NiO in red). All cells measured with
2
-
-
the same I /I electrolyte. Dotted lines are assigned to dark curves.
3
4
. Conclusion
In this contribution, two new push pull sensitizers PP1, PP2 and the associated dyad PP2-
NDI were specially designed for pDSC, synthesized and completely characterized by
spectroscopic and electrochemical means. The acquired experimental data confirmed that all
three dyes show excellent light harvesting abilities between 400 and 600 nm and are capable
to inject holes in the valence band of NiO. Significant photovoltaic performances of pDSC
fabricated with those new dyes with three very different electrolytes (triiodide/iodide,
-
3+
2+
disulfide/thiolate T /T and cobalt based Co1 /Co1 ) were obtained, showing the versatility
2
of the former. The best performances were nevertheless attained with PP2-NDI in
-
-
combination with the I /I electrolyte, which prompted us to build tandem cells allowing us to
3
obtain a tDSC which behaved well under irradiation, featuring the expected enhanced Voc of
ca. 900 mV and a satisfyingly high Jsc of 5 mA/cm² affording a high PCE of 2.8%, one of the
best ever reached by a tDSC. Moreover, both PP1 and PP2 feature quite negative reduction
potentials, providing sufficient regeneration driving force for a large panel of redox
mediators, and a diversity of catalysts for water or CO reduction in order to build dye-
2
8
8-91,92,93
sensitized photoelectrosynthesis cells.
5
. Acknowledgments
The authors are grateful to Agence Nationale de la Recherche (ANR) for the financial support
of these researches through the program POSITIF (ANR-12-PRGE-0016-01) and Région des
Pays de la Loire for the project LUMOMAT and Europe for COST CM1202 program
(
PERSPECT H O). A. F. acknowledges the European Research Council (ERC, Marches –
2
2
78845) for supporting his post-doctoral grant in Nantes. D. J. acknowledges the ERC for
financial support in the framework of the Marches Grant (278845). This research used
computational resources of CCIPL (Centre de Calcul Intensif des Pays de Loire), and a local
Troy cluster.
2
0

ACCEPTED MANUSCRIPT
6
. Experimental section
1
13
31
H, C and P NMR spectra were recorded on an AVANCE 300 UltraShield BRUKER and
1
13
AVANCE 400 BRUKER. Chemical shifts for H and C NMR spectra are referenced relative
1
to residual protium in the deuterated solvent (CDCl δ = 7.26 ppm for H and δ = 77.16 ppm
3
1
3
8
1
13
for C; THF-d δ = 3.57, 1.72 ppm for H and δ = 67.21, 25.31 ppm for C; CD OD δ = 3.31
3
1
13
ppm for H and δ = 49.00 ppm for C) or to an internal reference (TMS, δ = 0 ppm for both
1
13
H and C). NMR spectra were recorded at room temperature, chemical shifts are written in
ppm and coupling constants in Hz. High-resolution mass (HR-MS) spectra were obtained
either by electrospray ionization coupled with high resolution ion trap orbitrap (LTQ-
Orbitrap, ThermoFisher Scientific,) or by MALDI-TOF-TOF (Autoflex III, Bruker), working
in ion-positive or ion-negative mode. Electrochemical measurements were made under an
argon atmosphere in the mixture CH Cl /DMF: 95/5 with 0.1 M Bu NPF . Cyclic
2
2
4
6
voltammetry experiments were performed by using
a
SP300 Bio-Logic
potentiostat/galvanostat. A standard three-electrode electrochemical cell was used. Potentials
were referred to a saturated calomel electrode as internal reference. All potentials are quoted
relative to SCE. The working electrode was a glassy carbon disk and the auxiliary electrode
.
-1
was a Pt wire. In all the experiments the scan rate was 100 mV s . UV-visible absorption
spectra were recorded on a Variant Cary 300, using 1 cm path length cells. Emission spectra
were recorded on a Fluoromax-4 Horiba Jobin Yvon spectrofluorimeter (1 cm quartz cells).
Chemicals were purchased from Sigma-Aldrich or Alfa Aesar and used as received. Thin-
layer chromatography (TLC) was performed on aluminium sheets precoated with Merck 5735
Kieselgel 60F . Column chromatography was carried out either with Merck 5735 Kieselgel
2
54
6
0F (0.040-0.063 mm mesh).
9
4
Compounds named di-tert-butyl 4,4'-azanediyldibenzoate (2), 4-bromo-7-(thiophen-2-
9
5
yl)benzo[c][1,2,5]thiadiazole
(5),
4-(5-bromothiophen-2-yl)-7-(thiophen-2-
9
6
97
yl)benzo[c][1,2,5]thiadiazole (8) and ethynyl-NDI (13) were prepared according to already
described methods published in the literature.
Compound 3
Compound 2 (670 mg, 0.18 mmol, 1 eq) and 2,7-dibromo-9,9-bis(2-ethylhexyl)-9H-fluorene
(1) (2.0 g, 3.7 mmol, 2 eq) were dissolved in dry toluene (20 mL). The solution was degassed
2
1

ACCEPTED MANUSCRIPT
and potassium tertbutoxide (310 mg, 2.7 mmol, 1.5 eq), Pd(dba) (40 mg, 0.073 mmol, 0.04
2
eq), XPhos (70 mg, 0.15 mmol, 0.04 eq) were added. The resulting suspension was degassed
and stirred for 20 hours at 100 °C. After completion of the reaction monitored by TLC, the
mixture was diluted with dichloromethane and water. The mixture was extracted with
dichloromethane (3x) and the organic layer was washed with brine (2x), dried over anhydrous
magnesium sulfate, filtered and concentrated under reduced pressure. The crude product was
purified by column chromatography on silica gel (eluent: dichloromethane / petroleum spirit
1
[
3/2, v/v]) to give compound 3 (0.49 g, yield of 32%) as a light-yellow oil. H NMR (300
MHz, CDCl , 25 °C), δ (ppm): 7.85 (d, J= 8.8 Hz, 4 H), 7.62-7.49 (m, 4 H), 7.11-7.03 (m,
3
1
3
6
2
1
2
H), 1.97-1.73 (m, 4 H), 1.58 (s, 18 H), 1.03-0.40 (m, 30 H). C NMR (75 MHz, CDCl ,
3
5 °C), δ (ppm): 165.4, 152.8, 152.1, 150.7, 145.3, 139.5, 130.7, 130.1, 127.2, 126.1, 125.8,
22.3, 122.2, 122.1, 120.9, 120.8, 120.4, 80.7, 55.3, 44.4, 34.7, 33.4, 28.6, 28.2, 27.3, 26.4,
+
3.9, 22.9, 22.7, 20.8, 17.5, 14.1, 10.6, 9.9. HRMS (MALDI-TOF) m/z: [M] calculated for
C H BrNO : 835.4170; found: 835.4141. ∆= 3.5ppm.
51
66
4
Compound 6
Compound 5 (804 mg, 2.70 mmol, 1 eq) and hexamethylditin (2.65 g, 8.1 mmol, 3 eq) were
dissolved in dry DME (25 mL). The solution was degassed and the catalyst Pd(PPh ) (156
3
4
mg, 0.13 mmol, 0.05 eq) was added. The resulting mixture was stirred for 48 hours at 90 °C.
After completion of the reaction, the solvent was removed by rotary evaporation. The crude
product was purified by column chromatography on silica gel (eluent: petrol spirit) to give
1
compound 6 (670 mg, yield of 66%) as a yellow solid. H NMR (300 MHz, CDCl ), δ (ppm):
3
8
.08 (dd, J= 1.1 Hz, J=3.7 Hz, 1 H), 7.77 (d, J= 6.8 Hz, 1 H), 7.72 (d, J= 6.8 Hz, 1 H), 7.43
(dd, J= 5.1 Hz, J= 1.1 Hz, 1 H), 7.20 (dd, J= 3.6 Hz, J= 1.1 Hz, 1 H), 0.46 (s, JSn-H= 27.4 Hz,
1
3
JSn-H= 28.6 Hz, 9 H). C NMR (75 MHz, CDCl ), δ (ppm): 161.0, 151.4, 139.7, 137.5, 137.3,
3
1
19
1
36.9, 127.9, 127.6, 127.2, 126.5, -7.1, -7.2, -8.9, -10.7, -10.7. Sn NMR (150 MHz, C D ,
6 6
+
2
5 °C), δ (ppm): -28.3. HRMS (ESI+) m/z: [M+H] calculated for C H N S Sn: 378.9689;
13 15 2 2
found: 378.9689. ∆= 0.0 ppm.
Compound 9
2
2

ACCEPTED MANUSCRIPT
In a sealed tube were introduced compound 8 (250 mg, 0.66 mmol, 1 eq) and Pd(PPh ) (38
3
4
mg, 0.033 mmol, 0.05 eq) under argon. Degassed toluene (10 mL) was injected via septum.
Hexamethylditin (0.41 mL, 2.0 mmol, 3 eq) was then injected and the mixture was stirred at
1
00 °C for 6 h. The cooled solution was concentrated under reduced pressure and the crude
product was purified by column chromatography on silica gel (eluent: petroleum
spirit/triethylamine [90/10, v/v]). The desired compound 9 was obtained as an orange solid
1
with a yield of 86% (264 mg). H NMR (300 MHz, CDCl , 25 °C), δ (ppm): 8.18 (d, J= 3.5
3
Hz, 1 H), 8.10 (dd, J= 3.7 Hz, J= 1.1 Hz, 1 H), 7.84 (s, 2 H), 7.44 (dd, J= 5.1 Hz, J= 1.1 Hz, 1
H), 7.30 (d, J= 3.5 Hz, 1 H), 7.20 (dd, J= 3.7 Hz, J= 5.1 Hz, 1 H), 0.45 (s, JSn-H= 27.6 Hz, JSn-
1
3
H= 28.9 Hz, 9 H). C NMR (75 MHz, CDCl , 25 °C), δ (ppm): 152.6, 145.0, 140.4, 139.5,
3
1
19
1
36.1, 128.5, 128.0, 127.4, 126.7, 126.1, 125.9, 125.7, -5.6, -5.7, -8.1, -10.5, -10.6.
Sn
+
NMR (150 MHz, C D , 25 °C), δ (ppm): -25.3. HRMS (ESI+) m/z: [M+Na] calculated for
6
6
+
C H N S SnNa: 486.9390; found: [M+Na] 486.9390. ∆=0.7 ppm.
17
16
2 3
Compound 10
In a Schlenk tube under argon were introduced: compound 3 (50 mg, 0.060 mmol, 1 eq),
compound 6 (34 mg, 0.090 mmol, 1.5 eq), cesium fluoride (18 mg, 0.12 mmol, 2 eq), copper
iodide (1 mg, 0.0048 mmol, 0.08 eq) and PdCl (0.4 mg, 0.0023 mmol, 0.04 eq). A solution of
2
P(tBu) (2.0 mg, 0.010 mmol, 0.16 eq) in freshly distilled and degassed DMF (2 mL) was
3
injected via septum. The resulting mixture was stirred at 50 °C for 18 h. After cooling to room
temperature, the solution was diluted with dichloromethane and water. The mixture was
extracted with dichloromethane (2x). The combined organic layer was washed with brine
(2x), dried over anhydrous magnesium sulfate, filtered and concentrated under reduced
pressure. The crude product was then purified by column chromatography on silica gel
(eluent: petroleum spirit/dichloromethane [9/1, v/v] to [3/7, v/v]). 31 mg of the desired
1
compound IV-18 were obtained as a yellow-orange solid (yield of 53%). H NMR (300 MHz,
CDCl , 25 °C), δ (ppm): 8.07 (m, 1 H), 7.91 (m, 3 H), 7.80 (d, J= 8.6 Hz, 4 H), 7.73 (d, J= 7.8
3
Hz, 1 H), 7.69 (d, J= 7.6 Hz, 1 H), 7.64 (dt, J= 1.6 Hz, J= 8.1 Hz, 1 H), 7.40 (d, J= 5.0 Hz, 1
H), 7.16 (dd, J= 3.7 Hz, J= 5.1 Hz, 1 H), 7.10 (d, J= 1.6 Hz, 1 H), 7.03 (m, 5 H), 2.00 (m, 2
H), 1.82 (m, 2 H), 1.51 (s, 18 H), 0.71-1.03 (m, 18 H), 0.67 (t, J= 7.2 Hz, 3 H), 0.44-0.56 (m,
2
3

ACCEPTED MANUSCRIPT
1
3
9
1
1
2
H). C NMR (75 MHz, CDCl , 25 °C), δ (ppm): 165.4, 154.1, 153,0 150.8, 145.1, 140.9,
3
39.6, 133.4, 128.3, 127.9, 127.6, 127.4, 126.6, 126.2, 126.1, 125.9, 125.8, 124.8, 122.4,
22.1, 121.1, 119.5, 80.7, 55.3, 44.8, 44.7, 34.7, 33.4, 28.6, 28.2, 27.3, 26.4, 23.9, 22.9, 22.7,
+
0.8, 17.5, 14.1, 10.6. HRMS (MALDI-TOF) m/z: [M]
calculated for
C H N O S : 973.4881; found: 973.4915. ∆= 3.5ppm.
61
71
3
4 2
Sensitizer PP1
In a round bottom flask containing 10 (31 mg, 0.032 mmol, 1 eq) under argon was added
dichloromethane (6 mL) and then trifluoroacetic acid (2 mL) slowly. The resulting solution
was then stirred for 4 h at room temperature. After completion of the reaction monitored by
TLC, the crude was concentrated under reduced pressure. The product was obtained as a
1
yellow-orange powder with a yield of 99% (29 mg). H NMR (300 MHz, THF-d8, 25 °C),
δ (ppm): 8.27 (m, 2 H), 8.11 (m, 2 H), 7.85-8.00 (m, 7 H), 7.60 (d, J= 4.9 Hz, 1 H), 7.36 (m, 1
H), 7.15-7.26 (m, 6 H), 2.18 (m, 2 H), 2.02 (m, 2 H), 1.09 (m, 2 H), 0.85-1.04 (m, 16 H), 0.79
1
3
(
t, J= 7.1 Hz, 3 H), 0.58-0.66 (m, 9 H). C NMR (75 MHz, THF-d8, 25 °C), δ (ppm): 166.2,
1
1
3
54.0, 153.1, 152.3, 151.0, 150.7, 145.4, 141.0, 139.3, 138.6, 135.5, 132.9, 130.9, 128.3,
27.7, 127.6, 127.4, 126.8, 126.1, 125.5, 125.0, 122.8, 122.1, 121.2, 119.4, 55.3, 44.6, 44.4,
4.7, 34.2, 33.6, 33.5, 29.7, 28.8, 28.1, 27.3, 26.4, 26.4, 22.9, 22.6, 13.6, 10.1, 9.4, 9.4.
+
HRMS (ESI+) m/z: [M+H] calculated for C H N O S : 862.3712; found: 862.3737.
53
56
3
4 2
∆=2.9ppm.
Compound 11
In a sealed tube under argon were introduced: compound 3 (150 mg, 0.180 mmol, 1 eq),
compound 9 (180 mg, 0.390 mmol, 2.2 eq) in toluene (10 mL). The solution was degassed
carefully and Pd(PPh ) (21 mg, 0.018 mmol, 0.1 eq) was added quickly. The resulting
3
4
mixture was stirred at 110 °C for 24 h. After cooling to room temperature the solution was
concentrated under reduced pressure. The crude product was then purified by column
chromatography on silica gel (eluent: petroleum spirit/dichloromethane [6/4, v/v] to [4/6,
1
v/v]). 130 mg of the desired compound 11 were obtained as a red solid (yield of 69%). H
NMR (400 MHz, CDCl , 25 °C), δ (ppm): 8.13 (m, 2 H), 7.88 (m, 6 H), 7.67 (m, 4 H), 7.44
3
2
4

ACCEPTED MANUSCRIPT
(m, 2 H), 7.20 (m, 1 H), 7.17 (m, 1 H), 7.10 (m, 5 H), 2.07 (m, 2 H), 1.89 (m, 2 H), 1.60 (s, 18
1
3
H), 1.05 (m, 2 H), 0.79-1.01 (m, 16 H), 0.76 (m, 3 H), 0.69 (m, 3 H), 0.58 (m, 6 H). C NMR
100 MHz, CDCl , 25 °C), δ (ppm): 165.4, 152.8, 152.7, 152.6, 151.5, 150.8, 146.4, 145.1,
(
3
1
1
2
40.6, 139.4, 138.3, 138.3, 132.4, 130.8, 128.8, 128.0, 127.5, 126.8, 126.0, 126.0, 125.8,
25.2, 125.1, 123.8, 122.3, 121.3, 121.0, 120.0, 80.7, 55.3, 44.7, 34.8, 34.3, 33.7, 33.6, 28.8,
+
8.2, 27.5, 26.5, 26.5, 23.0, 22.8, 14.2, 14.1, 10.7, 10.1. HRMS (ESI+) m/z: [M] calculated
for C H N O S : 1055.4758; found: 1055.4741. ∆= 1.6ppm.
6
5
73
3
4 3
Sensitizer PP2
In a round bottom flask containing 11 (39 mg, 0.037 mmol, 1 eq) under argon was added
dichloromethane (5 mL) and then trifluoroacetic acid (1 mL) slowly. The resulting solution
was then stirred for 5 h at room temperature. After completion of the reaction monitored by
TLC, the crude was concentrated under reduced pressure. The product PP2 was obtained as a
1
red powder with a yield of 99% (35 mg). H NMR (400 MHz, CDCl /MeOD [4/1], 25 °C),
3
δ (ppm): 8.02 (m, 2 H), 7.85 (d, J= 8.3 Hz, 4 H), 7.77 (s, 2 H), 7.59 (m, 4 H), 7.35 (m, 2 H),
6
3
1
1
1
2
.98-7.13 (m, 7 H), 1.97 (m, 2 H), 1.80 (m, 2 H), 0.68-0.99 (m, 18 H), 0.66 (m, 3 H), 0.56 (m,
1
3
H), 0.46 (m, 6 H). C NMR (100 MHz, CDCl /MeOD [4/1], 25 °C), δ (ppm): 168.5, 152.9,
3
52.6, 152.4, 151.4, 151.1, 146.3, 144.7, 140.4, 139.3, 138.5, 138.5, 138.2, 132.4, 131.2,
28.7, 127.9, 127.4, 126.7, 126.3, 126.0, 125.7, 125.1, 124.9, 124.1, 123.7, 122.9, 122.7,
22.1, 121.4, 121.2, 121.0, 119.9, 55.2, 44.6, 34.7, 34.2, 33.6, 33.5, 31.5, 29.6, 28.7, 28.1,
+
8.1, 27.4, 26.4, 26.3, 22.8, 22.6, 22.5, 13.9, 13.8, 10.5, 9.9, 9.9. HRMS (ESI+) m/z: [M]
calculated for C H N O S : 943.3506; found: 943.3494. ∆=1.3 ppm.
5
7
57
3
4 3
Compound 12
In a round bottom flask equiped with an addition funnel and containing compound 11 (50 mg,
.047 mmol, 1 eq) and mercury caproate (24 mg, 0.057mmol, 1.2 eq) under argon atmosphere
0
was added a mixture chloroform/acetic acid [95/5] (5 mL). The solution was stirred 1 h at
room temperature. The solution was then cooled to 0 °C with an ice bath and a solution of
iodine (14 mg, 0.057 mmol, 1.2 eq) in chloroform (10 mL) was added dropwise over 2 h.
After 30 min of additionnal stirring, dichloromethane was added and the solution was washed
2
5

ACCEPTED MANUSCRIPT
with a 10% aqueous solution of sodium hydrogenocarbonate (2x), a saturated aqueous
solution of sodium sulfite (2x), and water (2x). The organic layer was then dried over
anhydrous magnesium sulfate, filtered and concentrated under reduced pressure. The
compound 12 was obtained as a red solid without further purification (47 mg, yield of 84%).
1
H NMR (400 MHz, CDCl , 25 °C), δ (ppm): 8.14 (m, 1 H), 7.86 (m, 5 H), 7.79 (d, J= 7.7 Hz,
3
1
H), 7.68 (m, 5 H), 7.44 (m, 1 H), 7.34 (d, J= 3.9 Hz, 1 H), 7.15 (m, 1 H), 7.05-7.13 (m, 5
H), 2.05 (m, 2 H), 1.87 (m, 2 H), 1.60 (s, 18 H), 1.03 (m, 2 H), 0.78-0.97 (m, 16 H), 0.75 (m,
1
3
3
1
1
1
2
H), 0.67 (m, 3 H), 0.55 (m, 6 H). C NMR (100 MHz, CDCl , 25 °C), δ (ppm): 165.5,
3
52.9, 152.5, 151.6, 150.9, 146.8, 145.4, 145.2, 140.7, 138.4, 138.3, 138.2, 137.9, 132.4,
30.9, 129.2, 128.3, 126.5, 126.1, 125.9, 125.5, 125.2, 124.7, 124.0, 122.7, 122.4, 121.5,
21.4, 121.1, 120.1, 80.8, 76.3, 55.4, 44.8, 34.9, 34.4, 33.7, 33.6, 28.9, 28.4, 28.3, 27.6, 26.6,
+
6.5, 23.1, 22.9, 14.3, 14.2, 10.9, 10.2. HRMS (ESI+) m/z: [M] calculated for
C H IN O S : 1181.3724; found: 1181.3716. ∆=0.7 ppm.
65
72
3
4 3
Compound 14
In a Schlenk tube were introduced: iodinated compound 12 (40 mg, 0.034 mmol, 1 eq),
compound 13 (32 mg, 0.068 mmol, 2 eq), Pd(PPh ) (2 mg, 0.0017 mmol, 0.05 eq) and
3
4
copper iodide (0.6 mg, 0.0034 mmol, 0.1 eq). The tube was backfilled with argon (5 cycles)
and a degassed mixture of toluene /triethylamine [3/1, v/v] (8 mL) was injected via septum.
The resulting mixture was stirred at 70 °C for 3 h. After cooling to room temperature, the
solution was concentrated under reduced pressure. The crude product was then purified by
column chromatography on silica gel (eluent: petroleum spirit/dichloromethane [9/1, v/v] to
pure dichloromethane). The wanted compound 14 was obtained as a dark red solid with a
1
yield of 70% (31 mg). H NMR (300 MHz, CDCl , 25 °C), δ (ppm): 8.78 (s, 4 H), 8.16 (m, 1
3
H), 8.01 (d, J= 3.9 Hz, 1 H), 7.87 (m, 6 H), 7.74 (d, J= 8.6 Hz, 2 H), 7.68 (m, 4 H), 7.45 (m, 1
H), 7.40 (d, J= 4.0 Hz, 1 H), 7.35 (d, J= 8.6 Hz, 2 H), 7.15 (d, J= 1.5 Hz, 1 H), 7.09 (m, 5 H),
4
.20 (m, 2 H), 2.05 (m, 2 H), 1.88 (m, 2 H), 1.76 (m, 2 H), 1.59 (s, 18 H), 1.24-1.48 (m, 10
1
3
H), 0.50-1.10 (m, 33 H). C NMR (75 MHz, CDCl , 25 °C), δ (ppm): 165.4, 162.9, 162.7,
3
152.8, 152.5, 152.4, 151.5, 150.8, 146.8, 145.1, 141.1, 140.7, 138.1, 134.5, 133.2, 132.5,
131.5, 131.0, 130.8, 129.2, 128.8, 127.2, 127.1, 128.8, 126.5, 126.0, 125.9, 125.1, 124.8,
124.2, 124.0, 123.9, 122.3, 121.0, 120.0, 94.2, 84.3, 80.7, 55.2, 44.7, 41.1, 34.8, 34.3, 33.6,
31.8, 29.3, 29.2, 28.8, 28.3, 28.1, 27.5, 27.1, 26.5, 23.0, 22.8, 22.7, 14.2, 14.1, 14.1, 10.7,
2
6

ACCEPTED MANUSCRIPT
+
1
0.0. HRMS (MALDI-TOF) m/z: [M] calculated for C H N O S : 1531.6494; found:
9
5
97
5
8 3
1
531.6539. ∆= 2.9 ppm.
Sensitizer PP2-NDI
In a round bottom flask containing 14 (31 mg, 0.024 mmol, 1 eq) under argon was added dry
dichloromethane (3 mL) and then trifluoroacetic acid (3 mL) slowly. Afterwards, the resulting
solution was stirred for 5 h at room temperature. After completion of the reaction monitored
by TLC, the crude was concentrated under reduced pressure. The resulting product PP2-NDI
was obtained as a dark red solid with a yield of 96% (27mg). NMR signals were very broad
and ill resolved regardless the conditions; however, TLC analysis revealed that the hydrolysis
step was quantitative, and the identity of the final molecule was confirmed by HRMS: HRMS
2
-
(
ESI-) m/z: [M-2H] calculated for C H N O S : 1417.5091; found: 1417.5116. ∆=1.8 ppm.
87 79 5 8 3
Photovoltaic devices fabrication
Conductive glass substrates (F-doped SnO ) were purchased from Solaronix (TEC15, sheet
2
resistance 15 Ω/square). Conductive glass substrates were successively cleaned by sonication
in soapy water, then acidified ethanol for 10 min before being fired at 450 °C for 30 min.
Once cooled down to room temperature, a solution of nickel acetate (0.5 M) and ethanolamine
(0.5 M) in methoxyethanol was spin-coated on FTO substrates, the latter were then fired at
7
9
5
00°C for 30 minutes. As treated FTO plates were then screen printed with NiO using a
home-made paste. The NiO screen-printing paste was produced by preparing a slurry of 3 g of
NiO nanopowder (Inframat) suspended in distilled ethanol (10 mL) and ball-milled (500 rpm)
for 24 h. The resulting slurry was mixed in a round-bottom flask with 10 ml of 10 wt%
ethanolic ethyl cellulose (Sigma Aldrich) solution and terpineol (20 mL), followed by slow
ethanol removal by rotary evaporation. The dried films were calcined in air at 400 °C for 0.5
h. The electrodes were finally dipped for 30 minutes in a 20 mM nickel acetate and
ethanolamine (a few drops to insure solubility) solution at 60°C, rinsed in ethanol and
8
0,81
thermally treated at 120°C in air.
The prepared NiO electrodes were soaked while still hot
(
80 °C) in a 0.2 mM solution (CH Cl /EtOH: 1/1) for each dye during 16 h.
2
2
Electrolytes used are composed of: 30 mM I , 0.1 M guanidinium thiocyanate, 0.5 M 4-
2
tertbutylpyridine and 0.6 M N-propyl-N’-ethylmethylimidazolium iodide in acetonitrile for
2
7

ACCEPTED MANUSCRIPT
-
-
2+
3+
the I /I electrolyte, 0.1 M Co1 , 0.1 M Co1 and 0.1 M LiClO in propylene carbonate for
3
4
-
-
cobalt complex as redox shuttle and 0.3 M T , 0.9 M T and 0.1 M LiTFSI for T /T
2
2
electrolyte. The electrolyte used for tandem dye sensitized solar cells was 0.6 M N-propyl-N’-
ethylmethylimidazolium iodide, 0.5 M 4-tertbutylpyridine, 0.1 M guanidinium thiocyanate
and 0.015 M I in acetronitrile. Platinum counter electrodes were prepared by depositing a
2
few drops of an isopropanol solution of hexachloroplatinic acid in distilled isopropanol (2 mg
per mL) on FTO plates (TEC7, Solaronix). Substrates were then fired at 375°C for 30 mn.
The photocathode and the counter electrode were placed on top of each other and sealed using
a thin transparent film of Surlyn polymer (DuPont, 25 µm) as spacer. A drop of electrolyte
was introduced through a predrilled hole in the counter electrode by vacuum backfilling, the
2
hole was then sealed by a glass stopper with surlyn. The cell had an active area of 0.25 cm .
The current-voltage characteristics were determined by applying an external potential bias to
the cell and measuring the photocurrent using a Keithley model 2400 digital source meter.
The solar simulator is an Oriel Lamp calibrated to 100 mW/cm².
7
. References
(
(
1)
2)
Armaroli, N.; Balzani, V. Angewandte Chemie-International Edition 2007, 46, 52.
Jean, J.; Brown, P. R.; Jaffe, R. L.; Buonassisi, T.; Bulovic, V. Energy & Environmental Science
2
015, 8, 1200.
(
(
3)
4)
Hagfeldt, A.; Boschloo, G.; Sun, L.; Kloo, L.; Pettersson, H. Chem. Rev. 2010, 110, 6595.
Fakharuddin, A.; Jose, R.; Brown, T. M.; Fabregat-Santiago, F.; Bisquert, J. Energy &
Environmental Science 2014, 7, 3952.
(
5)
Gratzel, M. Accounts of chemical research 2009, 42, 1788.
(
6)
Ito, S.; Murakami, T. N.; Comte, P.; Liska, P.; Gratzel, C.; Nazeeruddin, M. K.; Gratzel, M. Thin
Solid Films 2008, 516, 4613.
(
7)
O'Regan, B.; Grätzel, M. Nature 1991, 353, 737.
(
8)
Kakiage, K.; Aoyama, Y.; Yano, T.; Oya, K.; Fujisawa, J.-i.; Hanaya, M. Chemical
Communications 2015, 51, 15894.
9) Morandeira, A.; Boschloo, G.; Hagfeldt, A.; Hammarström, L. The Journal of Physical
Chemistry C 2008, 112, 9530.
(
(
10)
Odobel, F.; Le Pleux, L.; Pellegrin, Y.; Blart, E. Accounts of chemical research 2010, 43,
1
063.
(
(
11)
12)
He, J.; Lindström, H.; Hagfeldt, A.; Lindquist, S.-E. J. Phys. Chem. B 1999, 103, 8940.
He, J.; Lindström, H.; Hagfeldt, A.; Lindquist, S.-E. Solar Energy Materials and Solar Cells
2
000, 62, 265.
(
13)
Nattestad, A.; Perera, I.; Spiccia, L. Journal of Photochemistry and Photobiology C-
Photochemistry Reviews 2016, 28, 44.
(
(
14)
15)
Odobel, F.; Pellegrin, Y. The Journal of Physical Chemistry Letters 2013, 4, 2551.
Gibson, E. A.; Smeigh, A. L.; Le Pleux, L.; Fortage, J.; Boschloo, G.; Blart, E.; Pellegrin, Y.;
Odobel, F.; Hagfeldt, A.; Hammarström, L. Angew. Chem. Int. Ed. 2009, 48, 4402.
16) Nattestad, A.; Mozer, A. J.; Fischer, M. K.; Cheng, Y. B.; Mishra, A.; Bauerle, P.; Bach, U.
Nature materials 2010, 9, 31.
17) Powar, S.; Bhargava, R.; Daeneke, T.; Götz, G.; Bäuerle, P.; Geiger, T.; Kuster, S.; Nüesch, F.
A.; Spiccia, L.; Bach, U. Electrochimica Acta 2015, 182, 458.
18) Wood, C. J.; Summers, G. H.; Gibson, E. A. Chemical Communications 2015, 51, 3915.
(
(
(
2
8