Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 132 (2014) 477484

Contents lists available at ScienceDirect

Spectrochimica Acta Part A: Molecular and

Biomolecular Spectroscopy
journal homepage: www.elsevier.com/locate/saa

Absorption spectra and photovoltaic characterization of chlorophyllins

as sensitizers for dye-sensitized solar cells
Giuseppe Calogero , Ilaria Citro, Cristina Crupi, Gaetano Di Marco
CNR-IPCF, Istituto per i Processi Chimico-Fisici, via F. Stagno DAlcontres 37, I-98158 Messina, Italy

h i g h l i g h t s g r a p h i c a l a b s t r a c t

 The inuence of the solvent polarity

on the dye absorption bands was
analysed.
 A study of the Soret and Qy bands of
chlorophyllins was presented.
 The effect of copper on
chlorophyllins sensitizer properties
was investigated.
 The dependence of solar cells
performances by electrolyte
composition was studied.

a r t i c l e i n f o a b s t r a c t

Article history: Dye-sensitized solar cells (DSSCs) based on Chlorine-e6 (Chl-e6), a Chlorophyll a derivative, and Chl-e6
Received 17 February 2014 containing Cu, have been investigated by carrying out incident photon to current efciency (IPCE) and
Received in revised form 23 April 2014 currentvoltage (IV) measurements. The effect of the metallic ion and the inuence of the solvent polar-
Accepted 30 April 2014
ity on the dye aggregation and their absorption bands have been analysed by performing electronic
Available online 16 May 2014
absorption measurements. The dependence of the photoelectrochemical parameters of these DSSCs on
the electrolyte by the addition of pyrimidine and/or pyrrole has been discussed in details. For the rst
Keywords:
time IV curves for a DSSC based on copper Chl-e6 dye have been shown and compared with Zn based
Chlorophyllins
Dye-sensitized solar cells
chlorophyllin. Furthermore, the performance of a Cu-Chl-e6 based DSSC has been deeply improved by
Chlorophyll derivatives a progressive optimization of the TiO2 multilayer photoanode overcoming the best data reported in lit-
Titanium oxide nanoparticles erature so far for this dye. Its worth to emphasize that, the analysis reported in this paper supplies very
Soret useful information which paves the way to further detailed studies turned to the employment of natural
Natural dyes pigments as sensitizers for solar cells.
2014 Elsevier B.V. All rights reserved.

Introduction
Nowadays, the energetic needs matching the preservation of
cleaned environment have progressively become a very important
issue for the humankind. Considering that sunlight is clean,
Corresponding author. Tel.: +39 09039762247; fax: +39 0903974130.
E-mail address: calogero@me.cnr.it (G. Calogero).
abundant and readily available in most regions on the Earths sur-
face, the direct conversion of solar energy into electricity appears
an attractive sustainable and environmentally friendly solution to
energetic problems. Several new technologies have been developed
for this purpose and, among all the devices proposed, dye-sensitized
solar cells (DSSCs) represent, since the historic paper published in
1991 by Gratzel and OReagan [1] and up to now, the most promising
systems for low cost photovoltaic energy production.

http://dx.doi.org/10.1016/j.saa.2014.04.196
1386-1425/ 2014 Elsevier B.V. All rights reserved.
478 G. Calogero et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 132 (2014) 477484
A DSSC is a device based on the sensitization of large energy
band-gap semiconductor by a dye sensitizer. Such solar cell is typi-
cally assembled by placing in tandem a photoanode (dye-sensitized
semiconductor lm usually based on TiO2), an electrolyte solution
containing a redox system (iodide/iodine couple is commonly used)
and a counter-electrode [24]. The complex working mechanism of
a DSSC is mainly managed by the dye which plays a key role in light
absorption and in the process of electron injection into semiconduc-
tor. In particular, to accomplish this crucial task, the sensitizer must
have essential design characteristics: having an anchoring group
(typically carboxylic or hydroxyl groups) to strongly bind onto the
semiconductor and a remarkable solar radiation harvesting ability
in a broad wavelength range from the visible to the near-IR region.
Furthermore, the lowest unoccupied molecular orbital (LUMO)
and the highest occupied molecular orbital (HOMO) of the dye
should have suitable energetic values. More precisely, the LUMO
must be sufciently high in energy respect to the semiconductor
conduction band edge (CBE) for efcient charge injection into the
TiO2. On the other hand, the HOMO must be sufciently low in
energy respect to the redox couple of the electrolytic mediator for
efcient regeneration of the oxidized dye. Finally, electron transfer
from the dye to the TiO2 must also be rapid in comparison with
the sensitizer decay to its ground state and the sensitizer ground
and excited states must be stable [5]. Considerable efforts have been
directed towards the research of the most performing dye and, over
the years, the hypothesis of employing Chlorophylls to sensitize
dye-sensitized solar cells has been put forward. This idea has been
supported by the ability of these photosynthetic materials to con-
vert the light into electricity and inspired by the DSSC mechanism
that, for some aspects, reminds the natural photosynthesis involving
light-energy absorption and charge separation [6].
Chlorophylls, which drive photosynthesis in green plants, are
metal-complexes, with high symmetry, consisting of a tetrapyrrol-
ic macrocycle coordinated to a central magnesium ion [7].
The porphyrin macrocycle is the core of chlorophyll structures
and the nature of the frontier orbital in porphyrin macrocycles has
been an active area of research for computational and spectroscopic
investigations. Ab initio methods (restricted Hartree Fock, local den-
sity functional theory) have shown that both free base and metal
(Mg(II), Cu(II), Zn(II), Ni(II)) containing porphyrins generally possess
two a1u and a2u orbitals as the HOMO and HOMO1 [8]. The electron
density in a1u type orbital is essentially localized on the Ca and Cb
atoms of the pyrrole rings, while in the a2u orbital is mainly found
on the pyrrole nitrogens and on the meso carbons [9].
As for the highest occupied orbitals, the lowest lying unoccu-
pied orbitals are also similar in metal free and metal containing
porphyrins: there are two nearly degenerate eg(p*) orbitals essen-
tially centred on the porphyrin macrocycle. The persistence of
a1u(p) and a2u(p) orbitals in diverse metal porphyrins can account
for the facile formation of a p radical cation that plays an important
role in the biochemical function of porphyrins. For example, the p
radical cation formation is central in the photo-oxidation of chloro-
phyll which is essentially a Mg containing porphyrin. The nature of
the frontier orbitals accounts also for the similar features in the
absorption spectra of porphyrin derivatives. These show two char-
acteristic pp* absorption regions, the weaker Qy band in the visi-
ble range (550700 nm) and the intense Soret band in the near UV
region (400450 nm) [9]. Currently there are some synthetic sensi-
tizers based on tetrapyrrolic macrocycle such as phorphyrins [10]
and phtalocianines [11] which exhibit high efciency as sensitizers
but their preparation needs complicated, tedious, expensive and
toxic purications. Differently, Chlorophylls and theirs derivatives
are environmentally friendly and cheap materials and could be
prepared in the respect of the environment. Among chlorophylls,
Chlorophyll a (Chl a) represents the most abundant and widespread
natural compound. The rst study on Chl a and its derivatives as
sensitizers for TiO2 nanocrystalline in DSSCs was performed in
1993 [12]. In this study the greater sensitizer ability of Chl a deriv-
atives respect to the pure untreated chlorophyll a was evidenced.
This result can be explained considering the low attitude of this
natural pigment to be efciently absorbed on TiO2 layer when dis-
solved in some solvent (ethanol, acetone, pyridine, etc.) due to the
presence a phytyl ester group which, conversely, is eliminated in
its derivatives. Furthermore, following studies have evidenced that
an appropriate substitution of the functional groups bonded to the
tetrapyrrolic macrocycle [1316,7,17] or of the hosted central
metal [18,19,7] represents a useful procedure to tailor the photo-
chemical and photophysical properties of chlorophyll a derivatives
(commonly grouped under the name of chlorophyllins a, such as
phorphyrins, pheophorbides and chlorins).
The majority of scientic studies into the application of chloro-
phyllins a as sensitizer in DSSCs have focused on Chlorin e6 (Chl-e6)
with and without a central metal [12,2023], because this dye can
be easily synthesized from its precursor molecule even under mild
reaction conditions.
The macrocycle of the magnesium Chl-e6 is based on the core
structure of natural chlorophyll a obtained by the alkaline hydroly-
sis of this pigment. During this process, the phytil ester bond of Chl
a is saponied and the cyclopentanone ring is opened with the
formation of two additional carboxyl groups. Pure Chl-e6 [23]
and Chl-e6 trisodium salt [20] sensitized solar cells have shown a
photoenergy conversion efciency of 0.4% and 0.73%. Recently, this
efciency was increased to 0.9% by applying a scattering nanopar-
ticles coating onto the TiO2 layer [17].
A further analysis has revealed improved values of the current
density and of the open circuit voltage for a solar cell sensitized
by Chl-e6 coordinated to zinc ion [24]. Moreover, in 1993, Kay
and Grtzel [12] evidenced that copper Chl-e6 is an excellent pho-
tosensitizer for TiO2 with a 70% of the incident photons converted
into electrons but no photocurrentphotovoltage study was per-
formed on this DSSC.
With the aim to provide a further piece of information about
metal substituted Chl-e6 based DSSC, in this paper electronic
absorption and photocurrent action spectra, together with photo-
currentphotovoltage curves for Chl-e6 and copper Chl-e6 solar
cells are presented and compared with the existent literature.
The inuence of the solvent polarity on the dye absorption bands
is analysed. Finally, the dependence of the DSSC performance on
the electrolyte pH, on the dye central metal, on the employment
of a white or mirror reector behind the cathode, on the TiO2 lm
thickness and on the presence of a scattering coating on this semi-
conductor layer are discussed in detail.
Experimental section
Materials
All the solvents and the other chemicals employed for the
experiments were reagent or spectrophotometric grade and were
used as received without any further purication. The conductive
glassy plate (FTO-glass, Fluorine-doped SnO2, sheet resistance
7 ohm/cm2), and the Surlyn foils (thickness: 25 lm) were pur-
chased by Solaronix SA, while Ti-nanoxide (T/SP) was purchased
by Dyesol. The investigated chlorophyllins, chlorin-e6 (B10) and
Copper chlorin-e6 (A10) (see Fig. 1), were obtained by chlorophyll
a according to a well known procedure [25].
Preparation of TiO2 anodes
The FTO-glass, used for the preparation of transparent elec-
trodes, was rst cleaned in a detergent and then washed with
 G. Calogero et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 132 (2014) 477484
Fig. 1. Structural formula of the investigated chlorophyllin-a derivatives.
water and ethanol. These conductive glassy plates were immersed
in 100 ml of TiCl4/water solution (40 mM) at 70 C for 30 min,
washed with water and ethanol and dried in oven at 80 C for
30 min. A TiO2 layer was deposited on the FTO glassy plates by
screen printing (frame with polyester bres 43 T). This procedure,
involving two steps (coating and drying at 125 C), was repeated
twice. The TiO2 coated plates were gradually heated up to 325 C.
Then the temperature was increased to 375 C in 5 min, and after-
wards to 500 C. The plates were sintered at this temperature for
30 min [26], nally cooled down to room temperature and cut into
rectangular pieces (area: 2 cm  1.5 cm). Afterwards, the TiO2 lm
was treated with 40 mM TiCl4 solution, following the procedure
previously described, rinsed with water and ethanol and, nally,
sinterized at 500 C. Each anode resulted in a transparent meso-
scopic oxide lm, with a thickness of 12 lm and a spot area of
0.196 cm2, deposited on a rectangular FTO-glass.
The photoanode of an A10 based DSSC, with a nal thickness of
16 lm, was produced by modifying the manufacturing procedure,
previously described, by repeating the screen printing deposition
for three times and spreading a coating of TiO2 large nanoparticles
(furnished by Dyesol, sized 150200 nm) as scattering layer on
titania substrate.
The titanium oxide lm employed for UVVis absorption exper-
iments was prepared by screen printing method, using a frame of
polyester bres with a mesh size of 77.48 in order to obtain a
transparent ultrathin TiO2 lm with an estimated thickness of
about 3 lm and an active area of 1 cm2. To avoid moisture absorp-
tion, all the prepared anodes were stored in oven at about 80 C
until their employment.
Dye solution and soaking process
Dye solutions were prepared dissolving 1 g of chlorophyllin
(A10 or B10) dry powder in 100 ml of different solvents (water,
methanol or ethanol). A further solution was prepared by adding
a large amount of deoxycolic acid (785 mg) to the ethanol solution.
This expedient, according to literature data [27], assists dye disso-
lution, suppresses the aggregation of dye molecules and favours its
absorption onto titanium oxide lm.
The dye solutions were protected from light exposition and
stored in a refrigerator at +5 C until their use. The photoanodes
were prepared by soaking the screen printed glasses overnight in
the different dye solutions, at room temperature in the dark. The
excess dye was removed by rinsing the photoanodes with the same
solvent employed in the dye solution. Finally, these photo-elec-
trodes were dried in oven at 40 C for short time.
DSSC assembly
Each counter-electrode consisted in a FTO-glass plate (area:
2 cm  2 cm) on which a hole (1.0 mm diameter) was drilled. The
479
perforated substrates were washed and cleaned with water and
ethanol in order to remove any residual of glass powder or organic
contaminants. The Pt catalyst was deposited on the conductive face
of the FTO glass by dropping H2PtCl6 solution (5 mM in isopropanol)
and by heating at 500 C for 30 min. The photoanode and the Pt
counter-electrode were assembled into a sandwich type arrange-
ment and sealed (using a thermopress) with a hot melt gasket made
of the Surlyn ionomer.
Some electrolytes were prepared by dissolving the redox cou-
ple, I/I2 (LiI 0.7 M and I2 0.05 M), in 3-methoxypropionitrile
(MPN) solvent with and without the addition of pyrimidine
0.05 M or pyrrole 0.1 M. A further electrolyte based on acetoni-
trile/valeronitrile (4:3, v/v) was prepared with the addition of LiI
0.8 M and I2 0.05 M.
A drop of electrolyte solution was put on the hole in the back of
the cathode. The electrolyte was introduced into the cell via back-
lling under vacuum. The cell was placed in a small vacuum cham-
ber to remove air inside. Exposing it again to ambient pressure
causes the electrolyte to be driven into the cell. Finally, the hole
was sealed by using another Surlyn lm and closing by a thin cover
glass. An additive sealing with epoxydic resins was applied to
avoid the leakage of solvent vapours out from the cell and insure
a longer temporal stability of the device.
Apparatus and measurements
The thickness of the oxide lm deposited on the photoanodes
was measured by using a DektakXT prolometer (Bruker)
equipped with a diamond-tipped stylus (radius of 2 lm) and
selecting a vertical scan range of 65 lm with 8.0 nm resolution, a
programmed scan length of 6000 lm and a stylus force of 1 mg.
Each measure was veried by acquiring different runs, starting
from different positions set by rotating or translating the sample.
The absorption spectra of the photo-electrodes were recorded
by a Perkin Elmer L20 UVVis spectrophotometer, in the k range
from 375 nm up to 800 nm, with a speed of 240 nm/min and a step
of 2 nm. Incident photon-to-current conversion efciency (IPCE)
and relative photoaction spectra of DSSCs were measured by an
IPCE station. The IPCE station was composed by a 150 Xenon Light
Source (model ASB-XE, Spectral Products), a Monochromator
(model CM110, Spectral Products) equipped with a slit set, a Si cal-
ibrated detector (model 818-UV, Newport), a Picoamperometer
(model 6487, Keithley) and an IPCE Solarena Software. The IPCE
curve as a function of the incident wavelength has been deter-
mined by applying the following relation:
IPCEk LHEk  APCE 1
where LHE (k) represents the fraction of photons of wavelength k
absorbed by the dye and APCE is the absorbed photon to current
conversion efciency. Furthermore, the APCE is calculated by the
equation:
APCEk Uinj  Uc 2
where Uinj and Uc are the quantum yield for electron injection and
the electron collection efciency respectively.
CurrentVoltage curves were acquired by a digital Keithley 236
multimeter connected to a PC and controlled by a homemade pro-
gram. The sunlight irradiation with incident power, Pin, of 100 mW/
cm2 was provided by a LOT-Oriel solar simulator (Model LS0100-
1000, 300 W Xe Arc lamp powered by LSN251 power supply
equipped with AM 1.5 lter). Incident irradiance was measured
with a Si-based pyranometer, LP PYRA 02 AC. The ll factor (FF),
the open-circuit voltage (Voc), the current density (Jsc) and the pho-
toenergy conversion efciency (g) were acquired as photoelectro-
chemical parameters. In particular, the FF was obtained by the
following equation:
480 G. Calogero et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 132 (2014) 477484
Imax  V max
FF
Isc  V oc
where Imax and Vmax are the photocurrent and the photovoltage for
maximum power output and Isc is the short-circuit photocurrent.
Furthermore, the g parameter was determined by the relation:
Isc  V oc  FF
g 4
Pin
The IV measurements were carried out by putting a black back-
ground at the back of the solar cell. A10 based DSSCs were also
tested in the presence of a white and a mirror reector.
Result and discussion
Absorption spectra
The absorption spectra of the A10 and B10 dissolved in water,
methanol, ethanol and ethanol with deoxicolic acid are reported
in Fig. 2. All samples show an absorption spectrum characterized
by the presence of both Soret and Qy peaks. Other minor bands
are also visible, among which the closest to Qy peak is known as
Qx. The wavelengths, k1max and k2max, associated to these two main
absorption maxima are listed in Table 1.
Its worth noting that the spectra of A10 and B10 dyes dissolved
in ethanol and deoxicholic acid well agree with previously
reported literature data [12]. A red shift of the Soret peak and a
blue shift of the Qy peak for B10 sample in water respect to the cor-
responding system dissolved in alcoholic solutions are clearly
observed. The red-shift of the B10 Soret peak can be ascribed to
the effect of solvent polarity on dye. As a matter of fact, a solvent
with high polarity, such water, energetically stabilizes the dis-
solved dye, lowering the energy values of its molecular orbitals.
In particular, it inuences most the LUMO level, reducing the
Fig. 2. Absorption spectra of A10 (solid line) and B10 (dotted line) chlorophyllins in water (a), methanol (b), ethanol (c) and ethanol with deoxicholic acid (d).
energy gap HOMOLUMO and, thus, causing a red shift of its
3
absorption peaks.
Otherwise, the blue-shift revealed for the B10 Qy peak can be
explained taking into account a dye aggregation process. According
to literature data [24], a stacked face-to-face chlorine or porphyrin
p-aggregation leads to a blue-shift, while side-by-side chlorine or
porphyrin p-aggregation leads to a red shift respect to the corre-
sponding disaggregated sensitizer in the same solvent. Further-
more, it has been shown that the presence of acid in solution
favour the disaggregation [24]. Considering that the addition of
deoxicholic acid to ethanol keeps the Qy position almost
unchanged respect to pure methanol or ethanol solvent, no aggre-
gation for B10 in alcoholic solutions is supposed to take place. This
result is in agreement with a previous study on ZnChl-e6 which has
evidenced no aggregation for this dye dissolved in methanol solu-
tion in the presence or absence of cholic acid [24]. Its worth to
note that only a slight solvent effect on both Soret and Qy peaks
is observed for A10 system since the d orbitals of Cu ion, increasing
the electronic delocalization of porphyrin ring, make chlorophyllin-
a less polarisable respect to the B10 molecule. Differently, in the
light of these ndings, a face-to-face aggregation of B10 molecules
in water is supposed to occur causing the blue-shift observed for
B10 Qy peak which completely covers the expected red-shift due
to the higher polarity of this solvent. Furthermore, Cu chlorophyl-
lin-a dissolved in water is not expected to aggregate since the free
d orbitals of copper, coordinating a water molecule, does not allow
any aggregation of A10 molecules.
In Fig. 3, the absorption spectra normalized to the intensity of
the Soret peak for A10 and B10 systems in methanol are plotted
together with the spectra of Chl-e6 and ZnChl-e6 [28] previously
shown by Amao and Komori [28]. A well agreement between the
absorption trend of B10 dye and of pure Chl-e6 reported in literature
is observed. The only difference is represented by a slight broaden-
ing of the Soret peak in its right tail revealed for the sensitizer B10,
 G. Calogero et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 132 (2014) 477484
Table 1
Wavelength positions k1max and k2max of the Soret and Qy absorption band respectively and the intensity ratio ISoret/IQy. The specic solvent used for each dye solution is also
reported.
Dye Solvent/substrate
Chlorophyllins A10 Water
Chlorophyllins B10 Water
Chlorophyllins A10 Methanol
Chlorophyllins B10 Methanol
Chlorophyllins A10 Ethanol
Chlorophyllins B10 Ethanol
Chlorophyllins A10 Ethanol/deoxicholic acid
Chlorophyllins B10 Ethanol/deoxicholic acid
Fig. 3. Normalized absorption spectra of methanol dispersed: pure chlorophyllin
Chl-e6 (short dashed line) [25], A10 (solid line), B10 (dotted line) and Zn-Chle6
(short dot-dashed line) [28].
probably caused by a not complete purication of this sample
respect to the corresponding dye investigated in literature. The Qy
wavelength position is strongly affected by the specic central
metal and shows a blue-shift in all metallic-chlorophyllins respect
to k2max value revealed for Chl-e6 dye. More precisely, by comparing
the spectra of A10 and Zn Chl-e6, decreasing Qy wavelengths with
increasing central metal electronegativity is observed. Its worth
to note that, to our knowledge, this is the rst time that such a trend
is observed for metal-substituted Chl-e6 dyes. This result is in
agreement with previous analysis on metal-substituted porphyrins
[29,30], bacteriochlorophyll-a [31] and chlorophyll-a [32] which
have evidenced that the central metal interacts electrostatically
with the N-centred p-electron densities modifying the energy lev-
els of the electronic transitions of these systems. In particular, the
insertion of a highly electronegative ion has a more signicant
effect on the absorption spectrum bringing to a more evident Qy
blue-shift. This result is in perfect agreement with a recent
DFT-TDDFT computational study on similar metal-chlorophyll a
derivatives [33] which has evidenced that the insertion of metallic
ions into these sensitizers implies smaller HOMOLUMO gap whose
value decreases by changing magnesium with zinc ion. According to
these ndings, since copper ion has a similar electronic congura-
tion but a higher electronegativity than zinc metal, the Qy of
Cu-chlorophyll a derivative is expected to show the higher blue-shift
as observed for the investigated A10 system.
The intensity ratio, ISoret/IQy, of the Soret to Qy peak is also
reported in Table 1 for all the investigated dye solutions. These val-
ues give information about the absorption related to one peak
respect to the other: the lower this value is, the higher Qy absorp-
tion takes place. Furthermore, considering that the extinction
molar coefcient related to the Soret peak is higher than that asso-
ciated to the Qy peak, an enhanced Qy intensity or, equivalently, a
481
k1max (nm) k2max (nm) ISoret/IQy
409 632 3.22
412 652 5.91
404 626 4.40
401 660 3.94
405 629 3.68
402 663 3.07
409 628 2.64
403 662 1.73
lower ISoret/IQy ratio, points out a greater dye solubility into a spe-
cic solvent. In this regard, the lowest ISoret/IQy, revealed for A10
and B10 dissolved in ethanol/deoxicholic acid solution, indicates
that just in this solvent the sensitizers are well solubilized and
characterized by a higher light harvesting ability in the Qy range
than in the Soret region. For this reason these solutions have been
considered as the best choice to sensitize the anode of a DSSC.
In Fig. 4 the absorption spectra of A10 and B10 ethanol/deoxi-
colic acid solutions absorbed onto transparent titanium dioxide
layer (photoanode) are shown. As usual, a slight red-shift (about
5 nm) of the Qy for both systems in the absorption spectra of
photoanodes respect to those of the corresponding solutions is
observed [34], testifying the anchoring of the dye onto the semi-
conductor oxide surface. A more evident shift to near-IR region
for Zn-Chle6/methanol [28] absorbed onto TiO2 in comparison to
the corresponding solution has been revealed and ascribed to a
side-by-side chlorine aggregation which takes place in absence of
an appropriate dispersing medium. For the same system, in the
presence of cholic acid, a reduced Qy red-shift has been evidenced.
According to this result, it can be supposed that A10 and B10 mol-
ecules preserve their disaggregated state also on the TiO2 lm if
previously dispersed in ethanol/deoxicholic acid solution.
Referring to the shape of the acquired spectra, an enhanced
absorption in the wavelength region between Soret and Qy peaks
is observed for A10-photoanode respect to the B10-electrode.
A particular attention has to be focused on the Qy peak since,
according to a recent study [13], the absorption in this wavelength
region is at the origin of the photocurrent produced in chlorophyl-
lins sensitized solar cells. In fact, as a result of the Soret and Qx
energy slightly greater than the Qy state, these higher excited
states must be too short-lived to efciently inject electrons into
TiO2. Furthermore, differently from the Soret state which arises
from various transitions, the Qy-ground state transition is a well
dened single transition which should give rise to a specically
Fig. 4. Absorption spectra of A10 (solid line) and B10 (dotted line) chlorophyllins
absorbed onto TiO2 coated FTO-glass and the corresponding Lorentzian t of the Qy
peak (dashed line).
482 G. Calogero et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 132 (2014) 477484
oriented transitiondipole moment favouring the injection process
[35,36].
By carrying out a Lorentzian t procedure for the Qy peak in
both systems, an integrated area of 30% and 22% over the whole
absorption range (integrated area under the entire spectrum) is
determined for A10 and B10 systems respectively. This nding
underlines the higher light harvesting ability of the former pho-
toanode respect to the latter just in the Qy wavelength region
involved into the electron injection process between the dye and
TiO2 lm.
Photocurrent action spectra
In Fig. 5 the photoaction spectra of the DSSCs sensitized by the
two dyes, are reported. For both solar cells the IPCE of the Soret
peak is higher than that of the Qy, showing the same behaviour
observed in the corresponding dye absorption spectrum. By com-
paring the IPCE value of these two peaks for each pigment, some
considerations can be put forward. In agreement with a previous
investigation on chlorophyllins sensitized solar cells which has
proposed a model of electron injection by the Qy state [13], the
Soret state is expected to internally convert to the Qy state [14],
thus, the same injection yield for these two states can be supposed.
Furthermore, since the collection yield can be considered the same
for both peaks, depending only on cell assembly, material substrate
and TiO2-dye electron recombination process [37], the same APCE
for these two peaks is expected, according to Eq. (2). As a conse-
quence, the higher IPCE of the Soret peak can be ascribed to a
higher LHE respect to the Qy in agreement with the observed
absorption spectrum (see Fig. 4). Its worth to note that, the inter-
nal conversion of the Soret into the Qy is associated to an energy
loss which would reduce the DSSC efciency and whose extent is
proportional to the energy difference between these two states.
Following this statement, a rule can be proposed: the greater the
Soret-Qy wavelength difference is, the lower DSSC efciency is
expected. Obviously, the validity of this rule persists if dyes with
similar chemical-physical properties are considered.
The comparison of the two photoaction spectra of Fig. 5, points
out an increased IPCE of both Soret and Qy peaks due to the pres-
ence of Cu. The higher Qy IPCE value for A10-DSSC respect to the
B10-DSSC is in perfect agreement to the higher light harvesting
ability of the A10 respect to the B10 photoanodes in this wave-
length region as reported in the previous section. Conversely, an
opposite trend is observed for the Soret intensity of the two pig-
ments in the IPCE curves respect to the absorption spectra. Consid-
ering the higher absorption of B10 photoanode in the low
wavelength region respect to the A10 electrode, the lower Soret
Fig. 5. Photoaction spectra (IPCE curves) of DSSCs sensitized by A10 (solid line) or
B10 (dotted line) chlorophyllins.
IPCE value of the Chl-e6 respect to copper Chl-e6 based solar cells
can be explained in terms of a reduced APCE. At this stage no fur-
ther considerations can be put forward about Uinj and Uc relative
to the Soret region. The quantication of these two contributions
to the APCE would require a more detailed photo-physical analysis
similar to that previously reported for avonoid anthocyanin dyes
[37] and it will represent the main subject of an our future study.
In this regard, detailed investigations based on sub-nanosecond
laser spectroscopy [9] have demonstrated that non uorescent
chlorophyll derivatives like Cu chlorins have smaller charge injec-
tion rate constants (kinj = 3  108 s1) than dyes injecting from sin-
glet states like the free base chlorin e6 (kinj = 2.2  109 s1),
however the longer lived triplet state of Cu chlorins still allows
for a high electron injection quantum yield. Moreover competitive
excited state quenching due to exciton migration and annihilation
is reduced with metal porphyrins undergoing efcient intersystem
crossing, since long range energy transfer (Frster type) is spin
forbidden.[38]
Photocurrentphotovoltage characterization
In Fig. 6 the IV curves for A10 and B10 based DSSCs, containing
MPN as electrolyte, are shown and the corresponding photoelect-
rochemical parameters are reported in Table 2.
The comparison of these parameters highlights that, when a Cu
ion is introduced into chlorophyllin macrocycle, a better DSSC per-
formance is obtained with a nearly twofold solar energy to electric-
ity conversion efciency, g, respect to B10-based cell. Clearly, the
central metal strongly inuences both the density current, Jsc,
and the open circuit voltage, Voc, whereas it keeps the ll factor
almost unchanged. The increased Jsc value of A10-DSSC respect to
B10-DSSC one, is in perfect agreement with the IPCE results.
Concerning the Voc, a recent study on natural porphyrins and
chlorins [22] has evidenced an inverse proportionality between
the open circuit voltage and the difference of energy, DE, between
the dye HOMO level and the TiO2 conduction band edge (CBE), a
larger DE should lead to a smaller Voc. According to this result,
the lower Voc measured for A10-based DSSC respect to the value,
459 mV, reported in literature for ZnChle6-DSSC [24], should arise
from a higher DE value in agreement with a lower energy HOMO
level supposed for copper atom which is more electronegative than
zinc atom. Following this reasoning, the Voc of B10-DSSC would be
higher than A10-DSSC one, in contrast with the experimental
Fig. 6. Photocurrentphotovoltage curves of A10 (a) and B10 (b) solar cells with
MPN as electrolyte and black reector. In panel (a) the IV prole of improved A10
based DSSC (-d-) is also reported. In the inset the loading resistance as a function of
the generated power for metal substituted CHl-e6 DSSCs containing: MPN (-h-),
MPN with pyrrole 0.1 M (-D-) and MPN with pyrimidine 0.05 M (-s-) as
electrolytes.
 G. Calogero et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 132 (2014) 477484 483

Table 2
Principal parameters of IV measurements: current density Jsc, Fill factor FF, open circuit voltage Voc and efciency g.

Sample Electrolyte Jsc (mA/cm2) FF Voc (mV) g (%)

Chlorophyllins A10 MPN 3.36 0.65 400 0.88
Chlorophyllins A10 MPN/pyrrole 0.1 M 2.26 0.58 418 0.55
Chlorophyllins A10 MPN/pyrimidine 0.05 M 1.64 0.6 430 0.43
Chlorophyllins B10 MPN 1.88 0.67 357 0.45
Chlorophyllins B10 MPN/pyrrole 0.1 M 1.38 0.65 397 0.36
Chlorophyllins B10 MPN/pyrimidine 0.05 M 0.92 0.7 415 0.27

Table 3
Current density Jsc, Fill factor FF, open circuit voltage Voc and efciency g obtained by IV measurements on Cu-Chle6 based DSSC performed in the presence of a back background,
a white or a mirror reector.

Sample Electrolyte Jsc (mA/cm2) FF Voc (mV) g (%)

Chlorophyllins A10 black reector MPN 3.36 0.65 400 0.88
Chlorophyllins A10 white reector MPN 4.14 0.64 431 1.15
Chlorophyllins A10 mirror reector MPN 5.23 0.61 412 1.33

result reported in this paper. This apparent contradiction is over-
come considering that the higher Voc value of Cu sensitized solar
cell can be associated to an augmented ux of charge resulted from
injection, Iinj, according to the relation [39]:
kT Iinj
V oc ln
e nCB ket jI3 j
where k is the Boltzmann constant, T the temperature, e the elec-
tron charge, nCB the concentration of electrons at the TiO2 surface
and ket the constant rate of the triiodide reduction by the conduc-
tion band electrons of the titania at the semiconductor-electrolyte
interface [39,2]. The occurrence of this last process is supposed to
be almost the same for both A10 and B10 based DSSCs since strictly
related to the triiodide characteristic. In fact, this reduction is
favoured by the relatively small size of this ion which can either
cross the dye layer or penetrate nanometer-sized pores of the
TiO2 surface, which are not covered by the sensitizer and exposed
to the redox couple [40]. In agreement with this relation, the
reduced Iinj in B10-DSSC leads to the decrease of the open circuit
voltage. Moreover, this effect hides the Voc increase due to the lower
DE of B10 dye.
A further photocurrentphotovoltage study has been performed
on A-10 and B-10 solar cells by adding pyrrole or pyrimidine to the
MPN electrolyte and the obtained parameters are shown in Table 2.
All the photoelectrochemical parameters show higher values for
A10 based DSSC respect to B10-DSSC, independently from the spe-
cic electrolyte used. Its worth to note that, similarly, with the
presence of electrolyte additives, decreased current density and g
values, and increased open circuit voltage are observed respect to
the values determined either for A-10 or for B-10 based DSSCs con-
taining only pure MPN. According to previous studies [39], this
result can be ascribed to the effect of basic compound on the semi-
conductor conduction band edge. In fact, a shift of the semiconduc-
tor CBE to negative values has been revealed with increasing
electrolyte pH [39]. This occurrence would reduce the injection
from Qy state to TiO2 conduction band causing the observed
decreased Jsc value for both A10 and B10 solar cells, more clearly
evidenced in the presence of pyrimidine than with the addition
of pyrrole. Furthermore, the observed Voc increase is related to a
minor triiodide reduction process following the covering of TiO2
nanoparticles free space by the absorbed pyrrole or pyrimidine
as previously observed for other electrolyte additives [39,40].
In the insets of Fig. 6 the loading resistance as a function of gen-
erated power for DSSCs based on A10 and B10 sensitizers, contain-
ing different electrolytes, are shown. The addition of pyrimidine or
pyrrole to MPN electrolyte produces a progressive enhancement of
the loading resistance for both the investigated systems, probably
as a consequence of the increasing electrolyte viscosity. Further-
more, this effect is constantly more pronounced for the DSSC based
on B10 than on A10, in agreement with the worse performance
revealed for the former respect to the latter device.
5
In the light of all these experimental results, A-10 based DSSC
with MPN electrolyte appears to be the best solar cell prototype.
For this reason it was further investigated by performing IV mea-
surements in the presence of a white and a mirror reector put at
the back of the cathode. The obtained photoelectrochemical
parameters, together with the corresponding values of the same
solar cell with the black background, are reported in Table 3.
The insertion of the white reector brings to a marked light
scattering which supplies increased Jsc, Voc and efciency values
but keeps quite unchanged the FF parameter respect to the black
background. The light scattering effect is enhanced by the mirror
reector and this gives rise to the highest efciency and current
density values. Finally, with the aim of improving the DSSC perfor-
mance, an optimization procedure for the solar cell components
was carried out by increasing the semiconductor thickness, spread-
ing a scattering layer on the TiO2 substrate and using an electrolyte
based on acetonitrile/valeronitrile solution. The acquired IV curve
is also reported in Fig. 6(a). According to previous studies on natu-
ral dyes [39] the increased photoanode thickness and the presence
of the scattering layer improve the Jsc that reaches the value of
6.32 mA/cm2 owing to a greater amount of adsorbed dye on the
titania layer and to the harvesting of the scattered light by the sen-
sitizer. Furthermore, the lower viscosity of the electrolyte and the
increased I- content reduce the internal resistances of the solar cell
providing the best DSSC performance, characterized by a FF of 0.65,
a Voc of 438 mV and an efciency g of 1.81%. Although, at present,
these acquired parameters are not still competitive, its worth to
emphasize that, to the best of our knowledge, this is the rst time
that a copper-Chle6 sensitized solar cell is photoelectrochemically
investigated and fully studied. Moreover, the analysis reported in
this paper supplies very useful information which paves the way
to further detailed studies turned to the employment of natural
pigments as sensitizers for solar cells.
Conclusion
In this paper, electronic absorption and photocurrent action
spectra, together with photocurrentphotovoltage curves for
Chl-e6 and copper Chl-e6 solar cells have been presented. By
484 G. Calogero et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 132 (2014) 477484
performing absorption measurements on dyes dissolved in water
and alcoholic solutions, some information about the inuence of
solvent polarity on the light harvesting ability and aggregation of
these pigments have been inferred. For the rst time a blue shift
of the Qy wavelength position with increasing central metal elec-
tronegativity in Chle6 dyes has been discussed by taking into
account literature data. The best solubilisation and absorption in
the Qy range than in the Soret region has been observed for A10
and B10 dissolved in ethanol/deoxicholic acid solution. Both dyes
preserve their disaggregated state also on the TiO2 lm if previ-
ously dispersed in ethanol/deoxicholic acid solution. The copper
Chle6 photoanode has shown a greater light harvesting ability
respect to B10 electrode just in the Qy wavelength region. The
increased IPCE of both Soret and Qy peaks due to the presence of
Cu well agrees with the best performance obtained for the A10-
DSSC respect to the B10-DSSC.
Photocurrentphotovoltage measurements performed with dif-
ferent electrolyte suggest the use of co-additives to increase the
open circuit voltage as a consequence of a minor triiodide reduc-
tion process. Finally, the best A10-DSSC performance was obtained
by increasing the TiO2 lm thickness and spreading a scattering
coating. Most importantly, a rule has been proposed: the greater
the Soret-Qy wavelength difference is, the lower DSSC efciency
is expected, if dyes with similar chemical-physical properties are
compared.
Acknowledgements
This research was supported by EFOR Project (Energia da
FOnti Rinnovabili, Iniziativa CNR per il Mezzogiorno L. 191/2009
art. 2 comma 44) and by SAGRO Project nanced by Regione Sic-
iliana (Assessorato Regionale Attivit Produttive progetti PO FESR
2007/2013 linea dintervento 4.1.1.1). Furthermore, we acknowl-
edged Programma Operativo Nazionale Ricerca e Competitivit
20072013 projects FOTORIDUCO2 (PON 01_02257). We grate-
fully acknowledge Giuseppe Spinella, Roberto Caruso and Giuseppe
Lup, technicians of Mechanical and electronic lab at IPCF-CNR, for
the technical assistance and for the fabrication of accessories and
components useful for some electrical characterizations and
measurements.
References
[1] B. ORegan, M. Gratzel, Nature 353 (1991) 737740.
[2] G. Calogero, A. Sinopoli, I. Citro, G. Di Marco, V. Petrov, A.M. Diniz, A.J. Parola, F.
Pina, Photochem. Photobiol. Sci. 12 (2013) 883894.
[3] P. Calandra, G. Calogero, A. Sinopoli, P.G. Gucciardi, Int. J. Photoenergy (2010)
(Article ID 109495).
[4] G. Calogero, P. Calandra, A. Irrera, A. Sinopoli, I. Citro, G. Di Marco, Energy
Environ. Sci. 4 (2011) 18381844.
[5] M. Gratzel, Inorg. Chem. 44 (2005) 68416851.
[6] G. Calogero, I. Citro, G. Di Marco, S. Armeli Minicante, M. Morabito, G.
Genovese, Spectrochim. Acta A 117 (2014) 702706.
[7] X.F. Wang, H. Tamiaki, Energy Environ. Sci. 3 (2010) 94106.
[8] G. Lowe, Electronic Structure of Heme Sites, in: E.I. Solomon, A.B.P. Lever (Eds.),
Inorganic Electronic Structure and Spectroscopy, vol. 2, J. Wiley & Sons, New
York, p. 451.
[9] G. Calogero, G. Di Marco, S. Caramori, S. Cazzanti, R. Argazzi, C.A. Bignozzi,
Energy Environ. Sci. 2 (2009) 11621172.
[10] L. Lu-Lin, E. Wei-Guang, Diau, Chem. Soc. Rev. 42 (2013) 291304.
[11] G. Zanotti, N. Angelini, S. Notarantonio, A.M. Paoletti, G. Pennesi, G. Rossi, A.
Lembo, D. Colonna, A. Di Carlo, A. Reale, T.M. Brown, G. Calogero, Int. J.
Photoenergy (2010) (Article ID 136807 11 pages).
[12] A. Kay, M. Grtzel, J. Phys. Chem. 7 (3) (1993) 62726277.
[13] X.F. Wang, Y. Koyama, H. Nagae, Y. Wada, S. Sasaki, H. Tamiaki, J. Phys. Chem. C
112 (2008) 44184426.
[14] X.F. Wang, Y. Oyama, O. Kitao, Y. Wada, S. Sasaki, H. Tamiaki, H. Zhou, Biosens.
Bioelectron. 25 (2010) 19701976.
[15] X.F. Wang, H. Tamiaki, L. Wang, N. Tamai, O. Kitao, Ha. Zhou, S. Sasaki,
Langmuir 26 (2010) 63206327.
[16] X.F. Wang, O. Kitao, H. Zhou, H. Tamiaki, S. Sasaki, Chem. Commun. 12 (2009)
15231525.
[17] X.F. Wang, H. Tamiaki, O. Kitao, T. Ikeuchi, S. Sasaki, J. Power Sources 242
(2013) 860864.
[18] P.V. Kamat, J.P. Chauvet, R.W. Fessenden, J. Phys. Chem. 90 (1986) 13891394;
P.V. Kamat, J.P. Chauvet, Int. J. Radiat. Appl. Instrum. Part C, Radiat. Phys. Chem.
37 (1991) 705709.
[19] A. Singh, L.W. Johnson, Spectrochim. Acta A 58 (2002) 15731576.
[20] Y. Amao, T. Komori, Biosens. Bioelectron. 19 (2004) 843847.
[21] M. Ikegami, M. Ozeki, Y. Kijitori, T. Miyasaka, Electrochemistry 76 (2008) 140
143.
[22] X.F. Wang, O. Kitao, Molecules 17 (2012) 44844497.
[23] Y. Amao, Y. Yamada, K. Aoki, J. Photoch, Photobiol. A 164 (2004) 4751.
[24] Y. Amao, T. Komori, Biosens. Bioelectron. 22 (2007) 15611565.
[25] G. Oster, S.B. Broyde, J.S. Bellin, J. Am. Chem. Soc. 86 (1964) 13091313.
[26] S. Ito, P. Chen, P. Comte, M.K. Nazeeruddin, P. Liska, P. Pchy, M. Grtzel, Prog.
Photovoltaics Res. Appl. 15 (2007) 603612.
[27] Y. Amao, T. Komori, Langmuir 19 (21) (2003) 88728875.
[28] Y. Amao, Y. Yamada, Langmuir 21 (2005) 30083012.
[29] M. Gouterman, J. Mol. Spectrosc. 6 (1961) 138163;
M. Gouterman, in: D. Dolphin (Ed.), The Porphyrins, vol. 3, Academic Press,
New York, 1978, p. 1.
[30] J.H. Fuhrhop, K.M. Kadish, D.G. Davis, J. Am. Chem. Soc. 95 (1973) 51405147.
[31] G. Hartwich, L. Fiedor, I. Simonin, E. Cmiel, W. Schalfer, D. Noy, A. Scerz, H.
Scheer, J. Am. Chem. Soc. 120 (1998) 36753683.
[32] A. Drzewiecka-Matuszek, A. Skalma, A. Karocki, G. Stochel, L. Fiedor, J. Biol.
Inorg. Chem. 10 (2005) 453462.
[33] R. Rahimi, M. Rabbani, R. Rahimi, M.M. Waskasi, in: G.A. Seijas, M. Pilar
Vzquez-Tato (Eds.), Proceedings of the 14th International Electronic
Conference on Synthetic Organic Chemistry 130 November 2010, MDPI,
Basel, Switzerland, 2011, ISBN 3-906980-24-3.
[34] G. Calogero, G. Di Marco, Sol. Energy Mat. Sol. C 92 (2008) 13411346.
[35] J.D. Petke, G.M. Maggiora, L.L. Shipman, R.E. Christoffersen, Photochem.
Photobiol. 30 (1979) 203223.
[36] J.D. Petke, G.M. Maggiora, L.L. Shipman, R.E. Christoffersen, Photochem.
Photobiol. 32 (1980) 399414.
[37] N.J. Cherepy, G.P. Smestad, M. Grtzel, J.Z. Zhang, J. Phys. Chem. B 101 (1997)
93429351.
[38] A. Kay, R. Humphry-Baker, M. Graetzel, J. Phys. Chem. 98 (1994) 952959.
[39] M.K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-Baker, E. Mller, P. Liska, N.
Vlachopoulos, M. Grtzel, J. Am. Chem. Soc. 115 (1993) 63826390.
[40] G. Calogero, G. Di Marco, S. Cazzanti, S. Caramori, R. Argazzi, A. Di Carlo, C.A.
Bignozzi, Int. J. Mol. Sci. 11 (2010) 254267.