This document is the Accepted Manuscript version of a Published Work that appeared in final form in Dyes
and Pigments 126 (2016) 147e153 http://dx.doi.org/10.1016/j.dyepig.2015.11.002

Charge recombination losses in thiophene-substituted porphyrin dye-sensitized
solar cells
a

b

b

a

a

Susana Arrechea , John N. Clifford , Laia Pelleja , Ana Aljarilla , Pilar de la Cruz , Emilio
b, c, **
a, *
Palomares
, Fernando Langa
a
b
c

Universidad de Castilla-La Mancha, Instituto de Nanociencia, Nanotecnología y Materiales Moleculares (INAMOL), 45071, Toledo, Spain
Institute of Chemical Research of Catalonia (ICIQ), Avda. Països Catalans, 14, Tarragona, E-43007, Spain
ICREA, Passeig Lluis Companys, 23, Barcelona, E-08010, Spain

abstract
Two new porphyrins that incorporate thiophene substituents as spacers between the conjugated porphyrin core and the
anchoring cyanoacrylate group have been synthesised. The two dyes differ in the number of thiophene bridges; porphyrin 1a
has only one thiophene group and porphyrin 1b has two thiophene rings connected by a double bond. The measured light-toenergy conversion efficiencies in 1a and 1b were assessed using two different electrolytes, LP1 and LP2, which differ in the
presence of tert-butyl pyridine in LP1. An efficiency of 6% under standard measurement conditions has been achieved for 1a
þ LP1. However, for porphyrin 1b the use of electrolyte LP1 led to lower efficiencies and a value of approximately 4% was
obtained. The differences between the two types of solar cells and the electrolytes have been studied in-depth using photoinduced time-resolved techniques such as CE (Charge Extrac-tion) and TPV (Transient PhotoVoltage).

1. Introduction
Mesoporous TiO2 dye-sensitized photo-electrochemical solar cells, also
known as Gratzel€ solar cells, have achieved efficiencies as high as 12%
under sun-simulated conditions [1]. Recently, the first application of Building
Integrated Photo-Voltaic technology (BIPV) was demonstrated and this
showed the potential of a colourful and transparent solar cell technology [2].
In this type of solar cell the properties of the dye, such as colour, molecular
extinction coeffi-cient and robustness under illumination, are of utmost
importance. Among the examples of dyes published in the literature [3] for
Gratzel€ solar cells, porphyrins [4] have pleasant colours [5], excel-lent
molecular extinction coefficients [6,7] and good robustness [8,9]. Moreover,
based on the work of Yella and co-workers [1], efficiencies that surpass the
values obtained with ruthenium sen-sitizers have been achieved. However,
these remarkable

* Corresponding author.
** Corresponding author. Institute of Chemical Research of Catalonia (ICIQ), The Barcelona
Institute of Science and Technology, Avda. Països Catalans, 14, Tarragona, E-43007, Spain.
E-mail address: Fernando.Langa@uclm.es (F. Langa).

asymmetric porphyrins, which are based on DepeA molecular structures
(Donor-p-Acceptor), are difficult to synthesise and scale up to kilogram
quantities and this represents a major drawback for their use in solar modules
for BIPV.
Several groups, including our own, have proposed other types of
porphyrins [10e14] that are easier to synthesise and purify and show
efficiencies of around 5e7% under sun-simulated light irra-diation conditions
[15e20].
In the work described here we synthesised two novel Zn-porphyrin-based
structures (Fig. 1) that contain thiophene moi-eties as spacers between the Znporphyrin and the anchoring cyanoacrylic acid. The electronic properties of
the compounds were studied and they were also evaluated as components in
solar cells. The novel A3B Zn-porphyrins are tri-substituted with triphenylamine groups (TPA) as secondary electron donors (D); the volume of the
TPA groups also helps to diminish the natural tendency of porphyrins to
aggregate, which in turn diminishes the antenna effect of the porphyrins.

The efficiencies of the solar cells in two different electrolytes are
discussed and compared in terms of charge density, as measured by the
charge extraction technique, and device recombination kinetics under working
conditions, as measured using the photo-induced transient photovoltage. The
aim of the study was to demonstrate

Fig. 1. Structures of dyes 1a and 1b.

that the electrolyte interaction with the dye plays a critical role and this could
have a different effect on two porphyrins that differ only in the length of the
conjugated thiophene-based system used in the p-conjugated molecular bridge.

2. Results and discussion
The synthesis of dyes 1a,b was performed starting from trimethylsilylporphyrin 2 (Scheme 1) [10,21]. Firstly, the trimethylsilyl group was
quantitatively removed by reaction with TBAF and this was followed by
copper-free Pd-catalysed Sonogashira coupling [22] with 2-iodothiophene
derivatives 3aeb [23,24] Under these conditions, the reaction proceeded
smoothly to afford aldehydes 4a,b in 83% and 80% yield, respectively.
Subsequent Knoevenagel condensation of 4a,b with cyanoacetic acid, using
piperidine as base, yielded 1a,b in 89% and 82% yields, respectively, after
purifi-cation by column chromatography (silica gel, chloroform: methanol
1

13

10:1). All new compounds were fully characterised by FT-IR, H and C
NMR spectroscopies and by MALDI-TOF mass spectrometry (see
experimental section and ESI). It should be noted that in 4b and 1b the trans
character of the double bond was confirmed by the coupling constant of
1

around 15 Hz between the two vinyl protons in the H NMR spectra.
Compounds 1a,b were reasonably soluble in several common organic
solvents, such as CH2Cl2, CHCl3 and THF,

and this allowed the preparation of DSSCs. The thermal stabilities of
compounds 1a,b were evaluated by thermogravimetric analysis (TGA) under
nitrogen, with a heating rate of 10 C/min. The decomposition temperatures
(Td) were estimated from the TGA plots as the temperature of the intercept of
the leading edge of the weight loss curve. Under these conditions, compounds
1a and 1b display excellent thermal stability up to 200 C (Figs S17 and S18,
ESI) and this makes them suitable for application in photovoltaic devices.

2.1. Optical properties
The optical properties of 1a,b were studied by UV-Visible
spectrophotometry in solution and both compounds exhibited a panchromatic
absorption in the visible region. A solution of 1a in THF exhibited the
characteristic absorption pattern of a Zn-chelated porphyrin, with an intense
Soret band (B band) at 471 nm (log ε ¼ 4.99) and an intense intermolecular
charge transfer (ICT) band at 662 nm (log ε ¼ 4.53); between these bands,
one of the Q bands was observed at 583 nm (log ε ¼ 3.93). Extension of the
conjugation on the bridge by the introduction of a new thienyle-nevinylene
unit had a significant effect on the absorption spectrum of 1b, leading to a
panchromatic absorption as the Soret band was red-shifted to 460 nm (log ε ¼
5.21) and a new absorption band

Scheme 1. Synthetic route to dyes 1a and 1b.

appeared at 523 nm (log ε ¼ 4.87), which is attributed to the conjugated
oligomer bridge. Furthermore, the ICT band was observed at 668 nm (log ε ¼
4.91) with an increased absorption coefficient (see Fig. 2 and Table 1).
The steady state fluorescence spectra of dyes 1a,b measured in THF show
an emission band at 703 nm in both cases (1a lexc ¼ 471 nm and 1b lexc ¼
460 nm) (Fig. S19). It is also noteworthy that significant differences were not
observed in the emission band as a result of the increased conjugation due to the
inclusion of an additional thienylenevinylene unit. The emission bands were totally
quenched after adsorption onto TiO2 and this suggests an efficient photoinduced
electron transfer from the excited state of the dye to the TiO2 nanoparticles (Fig.
S20).

of the TPAs have a dihedral angle of around 42 . As a consequence, the
triphenylamine units act as bulky groups that make aggregation through p-pstacking of the dyes more difficult. This characteristic improves the
photoelectron injection efficiency from the dye to the TiO2 electrode and, as a
consequence, increases the conversion ef-ficiency. On the other hand, the
thiophene fragments of dyes 1a,b, which include the anchor group, lie in the
same plane as the porphyrin core. In dye 1b the planarity between the
macrocycle and the thiophene ring, which is bound to the triple bond, is
slightly greater (4 z 5 ) than that in dye 1a (4 z 8 ). This planarity be-tween
the porphyrin and the anchor group ensures effective elec-tron coupling
between the two electroactive fragments. The bond distances in the
conjugated systems based on thiophene are around 1.40 Å, both double and
single, and this provides evidence of the pushepull character and the effective
conjugation between the porphyrin and the anchor group.

2.2. Electrochemical properties
The redox properties of 1a,b were investigated by cyclic vol-tammetry
and square wave voltammetry in THF (Table 1, Figs S21 and S22). On the
cathodic side, compounds 1a,b show the first irreversible oxidation peaks at
þ

0.29 and 0.26 V (vs Fc/Fc ), respectively. This first oxidation potential is
assigned to the porphyrin core and the extended conjugation gives rise to a
decrease in the Eox value by 30 mV in 1b in comparison to 1a. The other
oxidation processes are attributed to the thienylenevinylene units. On the
reduction side, the two compounds show first reduction potentials at 1.62 V
and 1.71 V, respectively, as irre-versible waves, and these are attributed to the
reduction of the thienylenevinylene moieties. A second reduction potential
was also observed at 1.76 V for 1a and this is attributed to the reduction of
the porphyrin core. This second reduction potential was also observed in the
case of 1b as a shoulder on the first reduction wave. The EHOMO values
deduced from the oxidation potentials are 5.39 eV and 5.36 eV for 1a and 1b,
respectively, and these indicate that the regeneration is energetically feasible
e

by I/I 3 (Eredox ¼ 4.75 eV). The ELUMO values, e3.56 for 1a and 3.54 for
1b, also indicate that efficient electron injection into the TiO2 conduction
band (ETiO2 ¼ 4.00 eV) is energetically possible.

The molecular orbital analysis of the frontier orbitals of dyes 1a,b
together with their energy levels are shown in Fig. 4. The HOMO-LUMO
energy levels and consequently the HOMO-LUMO energy gap of dyes 1a
(1.99 eV) and 1b (1.88 eV) are very similar, but 1b has a somewhat more
stable HOMO energy level and a LUMO about 0.06 eV higher than that of
1a. Nevertheless, these types of sensi-tizers have sufficient driving force for
electron injection to TiO2.
The electron distributions of dyes 1a and 1b (Fig. 4) indicate that the
HOMO levels are mainly localized over the porphyrin macro-cycle with some
delocalization onto the nTV fragments and the anchor group. The HOMO-1
and HOMO-2 are confined to the tri-phenylamine units.
The first two LUMO orbitals are scattered along the porphyrin ring, the
thiophene units and the anchor group. The LUMOþ1 of dye 1a is exclusively
distributed on the porphyrin ring whereas some of the electron density of
LUMOþ1 of dye 1b is spread throughout the whole conjugated system, even
on the anchor group.
The above findings are consistent with the data obtained in the
1

electrochemical studies. The first oxidation potential (1a: E ox ¼ 0.29 V; 1b:
0.26 V) corresponds to the formation of the porphyrin radical cation and the
1

2.3. Theoretical calculations
In order to gain an insight into the geometrical and electronic properties of
dyes 1a and 1b, DFT calculations were performed at the B3LYP/6-31G(d)
level in vacuo with Gaussian 03W.
The calculated ground state geometries (Fig. 3) for both dyes (1a and 1b)
show that the meso-substituted aryl rings are twisted with respect to the
macrocyclic Zn-porphyrin core, with a dihedral angle of around 61e63 . The
phenyl groups bonded to the nitrogen atoms

second is due to the oxidation of the triphenylamino substituents (E ox z 0.5
V) in both dyes [10,21]. The reduction processes occur throughout the whole
conjugated system but in dye 1b the second reduction potential observed is
due to the reduction of the porphyrin macrocycle (see electro-chemical
studies).

2.4. Characterisation of the solar cells
Two different sets of porphyrin sensitized solar cells were pre-pared
depending on the electrolyte used. The electrolytes were

Table 1
a
a
b
UV-Vis, Fluorescence Emission and OSWV data for compounds 1aeb.

lmax/nm log ε lem/nm E
1a 662
583
471
305
1b 668
523
460
305
a

4.53
3.93
4.99
4.68
4.91
4.87
5.21
4.97
6

1

2
1
c
d
ox/V E ox/V E red/V E0-0/eV EHOMO/eV ELUMO/eV

703

0.29

0.44

1.62

1.83

5.39

3.56

703

0.26

0.49

1.71

1.82

5.36

3.54

6

1a: 8.70 10 M, THF; 1b: 5.15 10 M, THF.
3
[10
M] in THF versus Fc/Fcþ, glassy carbon, Pt counter electrode, 20 C, 0.1 M
1
Bu4NClO4, scan rate ¼ 100 mV s .
1
þ
c Calculated using the equation EHOMO (vs. vacuum) ¼ 5.1 e E ox (vs. Fc/Fc ) in eV [25].
b

Fig. 2. Normalised absorption spectra of dyes 1a (

6

) and 1b ( ) in THF solution (10 M).

d

ELUMO was calculated using ELUMO ¼ EHOMO þ E0-0, where E0-0 is the intersection of
the absorption and emission spectra.

Fig. 3. Structures of dyes 1a (left) and 1b (right) optimized at the B3LYP/6-31G(d) level.

denoted as LP1 and LP2 (see Experimental Section) and they differed only in
the presence of 0.5 M tert-butyl pyridine (TBP) in LP1. The current vs.
2

voltage (IV curves) under 100 mW/cm sun (1.5 AMG) simulated light are
shown in Fig. 5. As can be seen, there is a marked difference in the device
open circuit voltage (Voc) between the two porphyrin sensitized TiO2 solar
cells using the different electrolytes (Table 2).
Subsequent measurements aimed at characterising the solar cells were
carried out to analyse the differences in photocurrent. The IPCE (Incident
Photon to Current Efficiency) measurements (Fig. 6) are in good agreement
with the registered Jsc values under 1 sun (Table 2). As can be observed, in all
measurements two bands can be clearly differentiated and these correspond to
the Soret band of the porphyrins and the ICT band. Integration of the IPCE
spec-trum with respect to the 1.5 AM G solar spectrum gave the expected
2

photocurrent of 13 mA/cm

for 1a on using electrolyte LP1 and 12.8

2

Firstly, the charge density (total charge upon different light irradiation)
was measured using the photo-induced charge extrac-tion method, which has
been reported previously [12]. The different exponential curves obtained with
different light bias values (cell voltage upon different light illumination
intensities) are shown in Fig. 7. As one would expect, the use of 4-TBP with
the LP1 electrolyte leads to a shift in the measured curves towards higher Voc
e a finding in agreement with previous literature data [26]. The use of 4-TBP
induces an up-shift in the conduction band (CB) edge of the TiO2 and this
leads to an increase in cell voltage. Interestingly, for porphyrin 1a the shift in
the CB edge of the TiO2 did not lead to a concomitant decrease in the cell Jsc
value as a consequence of the more unfavourable electron transfer from the
dye excited state to the TiO2 CB, which is the case for porphyrin 1b. This
result led the authors to consider that the presence of 4-TBP, in combination
with porphyrin 1a, results in a better charge transfer from the dye excited
state due to a lower concentration of 1b molecular aggregates.

mA/cm for 1b on using electrolyte LP2.
Photo-induced charge recombination studies were carried out to analyse
the differences in Voc.

Measurement of the electron lifetimes for the different set of solar cells
(Fig. 8) shows that for solar cells sensitized with 1a the

Fig. 4. Energy levels of occupied and unoccupied MO and the electron density distribution of the corresponding frontier orbitals calculated at the B3LYP/6-31G(d) level.

2

Fig. 5. Current vs. Voltage measurements for the two different DSSC (cell area 0.16 cm ).
Dashed lines correspond to measurements made in the dark.

factor that has the most marked effect on solar cell performance is the use of
4-TBP in the electrolyte. It is noticeable that the presence of 4-TBP in the
LP1 electrolyte leads to slower charge recombina-tion kinetics between the
electrons at the TiO2 and the oxidised electrolyte when compared to the solar
cells based on 1a with the electrolyte LP2. In contrast, for porphyrin 1b the
presence of 4-TBP does not have an influence on the electron lifetime and in
both cases a similar value was obtained. Thus, the higher Voc obtained for
DSSC sensitized with 1b and the electrolyte LP1 is due to the up-shift of the

Fig. 6. IPCE measurements for the optimized DSSC.

finding that was attributed to the up-shift of the TiO2 CB leading to less
favourable charge transfer from the porphyrin excited state to the TiO2 CB. In
fact, the presence of 4-TBP did not slow the lifetime of the electrons in this
case.
4. Experimental
4.1. General procedure for the synthesis of 4a and 4b

TiO2 CB edge and not to slower charge combination kinetics between the
To

photo-injected electrons on the TiO2 and the oxidised electrolyte.

a

solution

of

[10,15,20-tri-(N,N-diphenylaniline)-5-

ethynyltrimethylsilane]porphyrinate zinc(II) 2 (1 mmol) in CH2Cl2 (200
mL/mmol) was added under argon TBAF (1.25 mmol, 1 M in THF). The
solution was stirred at room temperature for 1 h. The mixture was quenched
with H2O and extracted with CH2Cl2 (3 50 mL). The combined organic

3. Conclusions
Two related porphyrin sensitizers were designed and sensitized for use in
efficient DSSC. The two compounds differ in the number of thiophene
moieties. Moreover, the performance of these mate-rials was studied using
different electrolytes that differed only in the presence or absence of 4-TBP.
For porphyrin 1a the presence of 4-TBP in the electrolyte proved to be
beneficial, as reported previ-ously for other organic dyes. The expected shift
in the TiO2 con-duction band edge was observed and this led to a
concomitant increase in the solar cell Voc. Unexpectedly, an increase in the
solar cell photocurrent was also observed and this behaviour is not seen often
when using 4-TBP. The increase in the photocurrent was further confirmed by
measuring the IPCE. The logical explanation for the increase in photocurrent,
despite the up-shift in the TiO2 CB edge, is that in this particular case the
presence of 4-TBP led to the formation of fewer 1a molecular aggregates,
which in turn enhanced the electron transfer process from the excited state of

layers were dried over anhy-drous MgSO4 and the solvent was removed
under reduced pres-sure. The residue, which was used without further
purification, and 3aeb (3 mmol) were dissolved in dry THF (230 mL/mmol)
and Et3N (45 mL/mmol). The solution was degassed with argon for 15 min
and Pd2(dba)3 (0.3 mmol) and AsPh3 (2 mmol) were added to the mixture.
The solution was heated under reflux overnight. The sol-vent was removed
under reduced pressure. The product was pu-rified by column
chromatography (silica gel, hexane: CHCl3, 1:1).

4.2. Synthesis of 4a
[10,15,20-Tri-(N,N-diphenylaniline)-5-ethynyltrimethylsilane]
porphyrinate zinc(II) (110 mg, 0.09 mmol) was reacted according to the
general procedure to give 4a (107 mg) as a green solid

the molecule into the TiO2 CB.

In the case of porphyrin 1b, the presence of 4-TBP also led to an increase
in the solar cell Voc for the reason explained above but, in this case, a
decrease in the photocurrent was not observed e a

Table 2
Measured parameters for the optimised DSSC in Fig. 5.
2

Cell

Electrolyte

Sensitization
time (hours)

Jsc (mA/cm )

Voc (V)

FF (%)

h (%)

1a
1a
1b
1b

LP1
LP2
LP1
LP2

3
3
a
1
3

12.3
11.7
9.20
12.35

0.684
0.584
0.644
0.574

71
68
70
70

6.0
4.7
4.1
5.0

a

Longer sensitization times led to lower photocurrent and worse device per-formance. FF
(Fill Factor). h (solar-to-electrical current efficiency).

2

Fig. 7. Photo-induced charge extraction measurements for all DSSC (cell area 0.16 cm ).

4.4.1. Synthesis of [10,15,20-tri(N,N-diphenylaniline)-5-(2-cyano-3-(3,4dihexyl-5-ethynyl)thiophene-2-yl)acrylic acid)porphyrinate] zinc(II) (1a)
Compound 4a (60 mg, 0.043 mmol) was reacted according to
the general procedure to give 1a (56 mg) as a green solid (0.038 mmol, 89%
1

yield). H NMR (500 MHz, THF-d8) d/ppm: 9.72 (d, J ¼ 4.2 Hz, 2H), 9.06
(s broad, 2H), 8.94e8.90 (m, 4H), 8.54 (s, 1H), 8.05 (d, J ¼ 8.1 Hz, 6H), 7.45
(d, J ¼ 8.1 Hz, 6H), 7.40e7.39 (m, 24H), 7.13e7.96 (m, 6H), 3.29 (m, 2H),
3.00 (m, 2H), 2.08 (m, 2H),
1.77 (m, 2H), 1.61e1.50 (m, 6H), 1.47e1.41 (m, 6H), 0.98 (s broad, 3H), 0.87
13

(s broad, 3H). C NMR (125 MHz, THF-d8) d/ppm: 151.59, 150.75, 150.55,
148.82, 148.79, 148.24, 147.13, 137.78, 137.62, 136.24, 133.38, 132.47,
131.99, 130.47, 130.11, 125.48, 125.45, 124.01, 123.90,
123.85, 122.80, 121.99, 121.82, 32.84, 32.69, 32.57, 31.80, 31.55, 30.58,
Fig. 8. Electron lifetimes measured using TPV vs device charge density at different light bias
for all different DSSC.

1

30.44, 30.28, 23.51, 23.45, 14.34. FT-IR n/cm : 3419, 3068, 3033, 2955,
2925, 2155, 1600, 1495, 1386, 1317, 1282, 698. MS (m/z) (MALDI-TOF):
calculated for C96H78N8O2SZn: 1470.53; found: 1470.96. MP: 185e187 C.

1

(0.076 mmol, 83% yield). H NMR (400 MHz, CDCl3) d/ppm: 9.90 (s, 1H),
9.73 (d, J ¼ 4.7 Hz, 2H), 9.16 (d, J ¼ 4.7 Hz, 2H), 9.05 (d, J ¼ 4.7 Hz, 2H),
9.03 (d, J ¼ 4.7 Hz, 2H), 8.07 (t, J ¼ 8.7 Hz, 6H), 7.50e7.42 (m, 6H),
7.49e7.38 (m, 24H), 7.16e7.11 (m, 6H), 2.97 (m, 2H), 2.78 (m, 2H), 1.94 (m,
2H), 1.72e1.62 (m, 4H), 1.50e1.46 (m, 4H), 1.41e1.36 (m, 6H), 0.95 (t, J ¼
13

7.1 Hz, 3H), 0.88 (t, J ¼ 7.1 Hz, 3H). C NMR (100 MHz, CDCl3) d/ppm:
181.67, 152.00, 151.26, 151.16, 151.00, 147.87, 147.84, 147.69, 147.52,
137.18, 136.17, 136.00, 135.38, 135.34, 133.27, 132.44, 132.00, 130.35,
129.53, 124.91, 123.32, 122.45, 121.33, 121.26, 32.27, 31.86, 31.54, 30.95,
1

29.86, 29.39, 28.60, 22.81, 22.61, 14.15, 14.12. FT-IR n/cm : 2957, 2922,
2855, 1652, 1590, 1489, 1318, 1282, 996, 696. MS (m/z) (MALDI-TOF):
calculated for C93H77N7OSZn: 1403.52; found: 1403.90. MP: 180e182 C.

4.4.2. Synthesis of [10,15,20-tri(N,N-diphenylaniline)-5-(2-cyano3-(5-((E-2-(3,4-dihexyl-5-(ethynyl)thiophen-2-yl)vinyl)-3,4dihexylthiophene-2-yl)acrylic acid)porphyrinate] zinc(II) (1b)
Compound 4b (130 mg, 0.078 mmol) was reacted according to the
general procedure to give 1b (110 mg) as a green solid (0.063 mmol, 82%
1

yield). H NMR (500 MHz, THF-d8) d/ppm: 9.73 (d, J ¼ 4.4 Hz, 2H), 9.05
(d, J ¼ 4.4 Hz, 2H), 8.94e8.91 (m, 4H), 8.45 (s, 1H), 8.06 (t, J ¼ 8.6 Hz, 6H),
7.45 (t, J ¼ 8.6 Hz, 6H), 7.41e7.38 (m, 25H), 7.30 (d, J ¼ 15.4 Hz, 1H),
7.13e7.09 (m, 6H), 3.24 (m, 2H), 2.91 (m, 2H), 2.82 (m, 4H), 2.07 (m, 2H),
1.78 (m, 2H), 1.65 (m, 4H), 1.56 (m, 8H), 1.50e1.36 (m, 16H), 1.03e0.93 (m,
13

9H), 0.89 (t, J ¼ 7.2 Hz, 3H). C NMR (125 MHz, THF-d8) d/ppm: 151.45,
150.85, 150.47, 148.83, 148.80, 148.44, 148.21, 137.86, 137.74, 136.25,
136.22, 133.09, 132.36, 132.02, 130.48, 130.11, 125.46, 125.45, 123.87,
123.86, 122.55, 121.99, 121.85, 32.75, 32.50, 32.43, 32.30, 32.01, 31.90,
30.60, 30.45, 30.13, 30.07, 29.95, 23.55, 23.46, 23.40, 14.41, 14.37. FT-IR

4.3. Synthesis of 4b

1

[10,15,20-Tri-(N,N-diphenylaniline)-5-ethynyltrimethylsilane]
porphyrinate zinc(II) 2 (130 mg, 0.11 mmol) was reacted ac-cording to the
general procedure to give 4b (135 mg) as a green solid (0.080 mmol, 80%
1

yield). H NMR (400 MHz, CDCl3) d/ppm: 9.84 (s, 1H), 9.78 (d, J ¼ 4.6 Hz,
2H), 9.14 (d, J ¼ 4.6 Hz, 2H), 9.03 (dd, J ¼ 4.6, 6.7 Hz, 4H), 8.07 (dd, J ¼
7.6, 9.7 Hz, 6H), 7.49 (d, J ¼ 8.6 Hz, 6H), 7.45e7.41 (m, 24H), 7.32 (d, J ¼
15.5 Hz, 1H), 7.19e7.13 (m, 6H), 7.08 (d, J ¼ 15.5 Hz, 1H), 3.13 (dd, J ¼
6.0, 10.0 Hz, 2H), 2.77 (t, J ¼ 7.0 Hz, 2H), 2.74 (t, J ¼ 7.0 Hz, 2H), 2.61 (dd,
J ¼ 6.0, 10.0 Hz, 2H), 2.00 (q, J ¼ 7.0 Hz, 2H), 1.78e1.52 (m, 12H),
13

1.46e1.33 (m, 18H), 1.02e0.86 (m, 12H). C NMR (100 MHz, CDCl3)
d/ppm: 181.82, 152.97, 151.85, 150.78, 150.28, 149.90, 148.34, 147.89,
147.86, 147.46, 146.82, 142.77, 141.93, 137.17, 136.30, 136.18, 135.38,
134.78, 132.94, 138.28, 131.96, 130.47, 129.52, 124.86, 123.28, 122.91,
122.18, 121.40, 121.31, 119.24, 118.90, 32.30, 31.95, 31.71, 31.66, 31.57,
31.25, 29.93, 29.50, 29.42, 27.58, 27.12, 22.84, 22.72, 22.70, 22.60, 14.24,
1

14.18, 14.10. FT-IR n/cm : 3030, 2952, 2921, 2850, 1650, 1589, 1487, 1311,
1276, 995, 696. MS (m/z) (MALDI-TOF): calculated for
C111H105N7OS2Zn: 1679.71; found: 1680.10. MP: 177e178 C.

4.4. General procedure for Knoevenagel condensations
To a solution of 4aeb (1 mmol) in 20 mL/mmol of CHCl3 were added
cyanoacetic acid (1.5 mmol) and piperidine (0.1 mmol). The reaction mixture
was heated under reflux under argon for 18 h. The solvent was removed and
the residue was purified by column chromatography on silica gel with
CHCl3/MeOH 10:1 mixture as the eluent.

n/cm : 3389, 3059, 3032, 2950, 2932, 2924, 2855, 1595, 1564, 1491, 1386,
1277, 978, 796, 752, 691. MS (m/z) (MALDI-TOF): calculated for
C114H106N8O2S2Zn: 1746.72; found: 1747.17. MP: 182e184 ª C.

5. Device preparation and characterisation
The working electrodes for the best devices were made using 8 mm thick
transparent mesoporous TiO2 and 4 mm thick TiO2 layer, so-called 8 þ 4,
deposited onto fluorine-doped tin-oxide glass (FTO, Pilkington Glass Inc.,
with 15 Ohms/square sheet resistance). The counter electrode was fabricated
using the same FTO with a ther-malized Pt layer from H2PtCl6 (8% in water).
The subsequent steps to prepare the complete devices, as well as the
photocurrent vs voltage (IV curves), charge extraction and photo-induced
transient photovoltage have been published previously.
Photo-induced charge extraction was carried out as described in a
previous publication [26]. In brief, white light from a series of LEDs was used
as the light source. When the LEDs were turned off, the cell was immediately
short-circuited and the charge was extracted, thus allowing the electron
density in the cells to be calculated. By changing the intensity of the LEDs,
the electron density can be estimated as a function of cell voltage. In photoinduced transient photovoltage (TPV) measurements, in addition to the white
light applied by the LEDs, a diode pulse (660 nm, 10 mW) was applied to the
sample to induce a change of 2e3 mV within the cell. The resulting
photovoltage decay transients were collected and the t values were
determined by fitting the data to the equation exp( t/t).

5.1. Computational methods
The details for the computational methods can be found in the
supplementary information.
Acknowledgements
The ICIQ authors would like to thank MINECO for projects CTQ201347183 and the Severo Ochoa Excellence Accreditation 2014e2018 (SEV2013-0319). EP also thanks AGAUR for the SGR project 2014 SGR 763 and
ICIQ and ICREA for economical support. The UCLM authors would like to
thank MINECO for project CTQ2013-48252-P and Junta de Comunidades de
Castilla-La Man-cha (project PEII-2014-014-P). SA thanks to the Fundacion
Carolina for a grant.

Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.dyepig.2015.11.002.
References
[1] Yella A, Lee H-W, Tsao HN, Yi C, Chandiran AK, Nazeeruddin MK, et al. Science
2011;334:629e34.
[2] Joly D, Pelleja L, Narbey S, Oswald F, Meyer T, Kervella Y, et al. Energy Environ Sci
2015;8:2010e8.
[3] Clifford JN, Martinez-Ferrero E, Viterisi A, Palomares E. Chem Soc Rev
2011;40:1635e46.
[4] Martinez-Diaz MV, de la Torre G, Torres T. Chem Commun 2010;46: 7090e108.
[5] Kurotobi K, Toude Y, Kawamoto K, Fujimori Y, Ito S, Chabera P, et al. Chem e A Eur J
2013;19:17075e81.

[6] Mozer AJ, Griffith MJ, Tsekouras G, Wagner P, Wallace GG, Mori S, et al. J Am Chem
Soc 2009;131:15621e3.
[7] Chang Y-C, Wang C-L, Pan T-Y, Hong S-H, Lan C-M, Kuo H-H, et al. Chem Commun
2011;47:8910e2.
[8] Higashino T, Imahori H. Dalton Trans 2015;44:448e63.
[9] Hayashi S, Matsubara Y, Eu S, Hayashi H, Umeyama T, Matano Y, et al. Chem Lett
2008;37:846e7.
[10] Aljarilla A, Clifford JN, Pelleja L, Moncho A, Arrechea S, de la Cruz P, et al. J Mater
Chem A 2013;1:13640e7.
[11] Pelleja L, Kumar CV, Clifford JN, Palomares E. J Phys Chem C 2014;118: 16504e9.
[12] Cabau L, Vijay Kumar C, Moncho A, Clifford JN, Lopez N, Palomares E. Energy
Environ Sci 2015;8:1368e75.
[13] Barea EM, Caballero R, Lopez-Arroyo L, Guerrero A, de la Cruz P, Langa F, et al.
ChemPhysChem 2011;12:961e5.
[14] Pelado B, de la Cruz P, Gonzalez-Pedro V, Barea EM, Langa F. Tetrahedron Lett
2012;53:6665e9.
[15] Griffith MJ, Mozer AJ, Tsekouras G, Dong Y, Wagner P, Wagner K, et al. Appl Phys Lett
2011;98:163502.
[16] Armel V, Pringle JM, Forsyth M, MacFarlane DR, Officer DL, Wagner P. Chem
Commun 2010;46:3146e8.
[17] Lee C-W, Lu H-P, Lan C-M, Huang Y-L, Liang Y-R, Yen W-N, et al. Chem A Eur J
2009;15:1403e12.
[18] Griffith MJ, Sunahara K, Wagner P, Wagner K, Wallace GG, Officer DL, et al. Chem
Commun 2012;48:4145e62.
[19] Wu S-L, Lu H-P, Yu H-T, Chuang S-H, Chiu C-L, Lee C-W, et al. Energy Environ Sci
2010;3:949e55.
[20] Arteaga D, Cotta R, Ortiz A, Insuasty B, Martin N, Echegoyen L. Dyes Pigments
2015;112:127e37.
[21] Ragoussi M-E, de la Torre G, Torres T. Eur J Org Chem 2013;2013:2832e40.
[22] Wagner RW, Johnson TE, Li F, Lindsey JS. J Org Chem 1995;60:5266e73.
[23] Pelleja L, Dominguez R, Aljarilla A, Clifford JN, de la Cruz P, Langa F, et al.
ChemElectroChem 2014;1:1126e9.
[24] Arrechea S, Molina-Ontoria A, Aljarilla A, de la Cruz P, Langa F, Echegoyen L. Dyes
Pigments 2015;121:109e17.
[25] D'Andrade BW, Datta S, Forrest SR, Djurovich P, Polikarpov E, Thompson ME. Org
Electron 2005;6:11e20.
[26] Zewdu T, Clifford JN, Hernandez JP, Palomares E. Energy Environ Sci 2011;4: 4633e8.