Bis(1,10-phenanthroline) copper complexes with tailored molecular architecture: from electrochemical features to application as redox mediators in dye-sensitized solar cells

Article abstract of DOI:10.1016/j.electacta.2018.03.103

In the last few years, copper coordination compounds turned out to be effective competitors of cobalt complexes as redox mediators in the formulation of iodine-free electrolytes for dye-sensitized solar cells (DSSCs). However, the lack of a clear correlation between electrochemical signatures of copper complexes (i.e. half-wave potential and heterogeneous electron transfer rate) and photoelectrochemical performance of solar devices makes difficult the optimization of their coordination sphere. Therefore, to partially fill this gap and to elucidate the intrinsic correlation between the molecular architecture of these complexes and their electrochemical features, we prepared four Cu^+/2+ redox couples in which the copper center is coordinated by two 1,10-phenanthrolines bearing various substituents in position 2. These complexes were well characterized, from both electrochemical and spectroscopic point of view, and tested as electron shuttles in lab-scale photoelectrochemical cells sensitized with two efficient π-extended benzothiadiazole dyes. It appeared that 2-aryl-1,10-phenanthrolines effectively combine suitable optical and electrochemical properties. While a fast electron transfer kinetics generally positively affects the dye regeneration process, an optimal balance between dye regeneration efficiency, mass transport and heterogeneous electron transfer at both the counter electrode and at the TiO₂ interface, must be achieved in order to optimize DSSC performance. Within our series, the top performer was [Cu(2-tolyl-1,10-phenanthroline)₂]^+/2+ which achieved a relative 20% and 15% improvement in power conversion efficiency (under 100 mW s^?1 simulated AM 1.5G illumination) with respect to control cells filled with [Co(bpy)₃]^2+/3+ (bpy = 2,2′-bipyridine) and I^?/I₃ ^? electrolytes, respectively.

Full text of DOI:10.1016/j.electacta.2018.03.103

Accepted Manuscript
Bis(1,10-phenanthroline) copper complexes with tailored molecular architecture: From
electrochemical features to application as redox mediators in dye-sensitized solar
cells
Elisabetta Benazzi, Mirko Magni, Alessia Colombo, Claudia Dragonetti, Stefano
Caramori, Carlo Alberto Bignozzi, Roberto Grisorio, Gian Paolo Suranna, Maria Pia
Cipolla, Michele Manca, Dominique Roberto
PII:
S0013-4686(18)30605-4
10.1016/j.electacta.2018.03.103
EA 31469
DOI:
Reference:
To appear in: Electrochimica Acta
Received Date: 25 January 2018
Revised Date: 15 March 2018
Accepted Date: 17 March 2018
Please cite this article as: E. Benazzi, M. Magni, A. Colombo, C. Dragonetti, S. Caramori, C.A.
Bignozzi, R. Grisorio, G.P. Suranna, M.P. Cipolla, M. Manca, D. Roberto, Bis(1,10-phenanthroline)
copper complexes with tailored molecular architecture: From electrochemical features to application
as redox mediators in dye-sensitized solar cells, Electrochimica Acta (2018), doi: 10.1016/
j.electacta.2018.03.103.
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
Bis(1,10-phenanthroline) copper complexes with tailored molecular
architecture: from electrochemical features to application as redox mediators in
dye-sensitized solar cells.
Elisabetta Benazzi,a Mirko Magni,b* Alessia Colombo,b** Claudia Dragonetti,b Stefano Caramori,a
Carlo Alberto Bignozzi,a Roberto Grisorio,c Gian Paolo Suranna,c Maria Pia Cipolla,d,e Michele
Mancad*** and Dominique Robertob
a Dipartimento di Scienze Chimiche e Farmaceutiche, Università di Ferrara, Via F. di Mortara,
Ferrara, Italy.
b Dipartimento di Chimica, Università degli Studi di Milano, UdR dell’INSTM, Via Golgi 19,
Milano, Italy.
c Dip. di Ingegneria Civile, Ambientale, del Territorio, Edile e di Chimica, Politecnico di Bari, via
Orabona, Bari and CNR-NANOTEC c/o Campus Ecotekne, via Monteroni, Lecce, Italy.
d Center for Biomolecular Nanotechnology - Fondazione Istituto Italiano di Tecnologia, Via
Barsanti 14, Arnesano (Lecce), Italy.
e Dipartimento di Ingegneria dell' Innovazione, Università del Salento, Campus Ecotekne, Via per
Monteroni, 73100, Lecce, Italy.
Corresponding authors e-mail: *mirko.magni@unimi.it; **alessia.colombo@unimi.it;
***michele.manca@iit.it
Abstract
In the last few years, copper coordination compounds turned out to be effective competitors of
cobalt complexes as redox mediators in the formulation of iodine-free electrolytes for dye-
sensitized solar cells (DSSCs). However, the lack of a clear correlation between electrochemical
1

ACCEPTED MANUSCRIPT
signatures of copper complexes (i.e. half-wave potential and heterogeneous electron transfer rate)
and photoelectrochemical performance of solar devices makes difficult the optimization of their
coordination sphere. Therefore, to partially fill this gap and to elucidate the intrinsic correlation
between the molecular architecture of these complexes and their electrochemical features, we
prepared four Cu+/2+ redox couples in which the copper center is coordinated by two 1,10-
phenanthrolines bearing various substituents in position 2. These complexes were well
characterized, from both electrochemical and spectroscopic point of view, and tested as electron
shuttles in lab-scale photoelectrochemical cells sensitized with two efficient π-extended
benzothiadiazole dyes. It appeared that 2-aryl-1,10-phenanthrolines effectively combine suitable
optical and electrochemical properties. While a fast electron transfer kinetics generally positively
affects the dye regeneration process, an optimal balance between dye regeneration efficiency, mass
transport and heterogeneous electron transfer at both the counter electrode and at the TiO2 interface,
must be achieved in order to optimize DSSC performance. Within our series, the top performer was
[Cu(2-tolyl-1,10-phenanthroline)2]+/2+ which achieved a relative 20% and 15% improvement in
power conversion efficiency (under 100 mW s–1 simulated AM 1.5G illumination) with respect to
–
control cells filled with [Co(bpy)3]2+/3+ (bpy=2,2’-bipyridine) and I–/I3 electrolytes, respectively.
Keywords:
homoleptic bis(1,10-phenanthroline) copper complexes; copper redox mediators; iodine-free
electrolyte; dye-sensitized solar cells; structure-activity relationship
1. Introduction
Dye-sensitized solar cells (DSSCs), which are composed of mesoporous nanocrystalline TiO2 layers
modified with photosensitizing dyes as photoanodes, redox electrolytes and counter electrodes
2

ACCEPTED MANUSCRIPT
(Figure 1), have been actively investigated as photovoltaic devices in alternative to conventional
silicon-based inorganic solar cells, because of their potentially low production costs, relatively low
toxicity of the constituent elements and relatively high photon-to-current conversion efficiency
(PCE) especially under low-light intensities and scattered light conditions [1].
(Figure 1)
DSSCs, easy to fabricate on flexible panels, offer unique transparency and aesthetic advantages
with respect to building-integration. However, the promise of DSSCs to become the largest
commercially available solar energy conversion technology and the most convenient alternative to
silicon is still far to get fulfilled. Among the factors still obstructing the lab-to-market bridge
crossing road, there is the failure to achieve DSSCs with long-term stability. Also, further reducing
their costs remains an important target as the cost of silicon solar cells has significantly dropped in
recent years [2-3].
Since the milestone paper of O’Regan and Grätzel on DSSCs in 1991 [4], photosensitizers have
monopolized attention of researchers aimed to maximize their light harnessing, by improving the
overlap between the dye absorption spectrum and the solar emission one, and enhance their dipole
moment to facilitate the charge separation, by creating suitable conjugated moieties with a donor-π-
linker-acceptor architecture [5-10]. For many years, the most efficient DSSCs were based on
–
variously substituted polypyridyl ruthenium complexes as dye and the iodide/triiodide couple (I–/I3
) as electrolyte, reaching PCE around 10-12 % [11-13]. However, in the last decade, the high cost
and rarity of ruthenium encouraged the search for efficient noble-metal-free dyes such as organic
compounds [5], porphyrins [7], and copper complexes [8-10]. In parallel, an exponential growth of
work has been devoted to find alternative redox mediators such as metal transition complexes [14-
–
15] to replace the traditional redox couple I–/I3 which has many problems such as a too negative
redox potential, complicated two-electron redox chemistry, corrosiveness and volatility of I2 in
–
equilibrium with I3 complicating long-term cell sealing and therefore hampering long-term stability
[16-17]. The use of Co2+/3+ complexes as mediators has addressed some of these issues since they
3

ACCEPTED MANUSCRIPT
are outer-sphere one-electron redox systems with more positive and tunable redox potentials,
affording large Voc [16-17]. In fact, to date record power conversion efficiencies (13-14%) in
DSSCs have been reached with a cobalt-based electrolyte in combination with either porphyrins
[18-19] or π-delocalized organic dyes [20]. Nevertheless, cobalt complexes have some drawbacks
associated with slow mass transport in the electrolyte solution and large internal reorganization
energy between the high-spin d7 and low-spin d6 states. The limitations caused by the
reorganization energy are minimized by using well-designed Cu+/2+ complexes with distorted
tetragonal geometry [21]. These alternative electron shuttles are of particular interest since copper is
a low-cost, environmental friendly metal.
As recently reviewed [10], copper complexes were firstly proposed as electron shuttles in 2005
when, in combination with N719 as dye, [Cu(2,9-dimethyl-1,10-phenantholine)2]+/2+ reached a PCE
of 1.4%, ten-fold higher than that of related [Cu(1,10-phenantholine)2]+/2+, showing the crucial role
of the methyl groups in position α to the coordinating nitrogen atoms in order to maintain the
distorted tetragonal geometry around the copper center and therefore a very efficient electron
transfer process [22]. By using the same redox mediator with various Donor-π-Acceptor organic
dyes, there was a progressive growth of PCE from 7% [23] to 8.3 % [24] to 10.3% [25]. Besides,
the use of [Cu(1,1-bis(2-pyridyl)ethane)2]+/2+ [26] and [Cu(6,6’-dimethyl-2,2’-bipyridine)2]+/2+ [25,
27] redox mediators, in combination with organic dyes, led to PCE of 9.0% and 10.3%,
respectively. These copper-based mediators overcame performance of the control cells filled with a
reference cobalt-based electrolyte [24, 26,27]. Interestingly, [Cu(4,4’,6,6’-tetramethyl-2,2’-
bipyridine)2]+/2+ mediators revealed also very promising for indoor application [28]. Clearly the use
of copper-based mediators appears as a key in the fabrication of convenient low-cost DSSCs.
Remarkably, we found that the novel redox couple [Cu(2-mesityl-4,7-dimethyl-1,10-
phenantholine)2]+/2+ is fascinating as electron shuttle because the presence of the bulky mesityl
groups on the phenanthrolines protects as a kiss lock the copper center, leading to slightly superior
dye regeneration kinetics and longer electron lifetimes with respect to simpler 2,9-dimethyl-1,10-
4

ACCEPTED MANUSCRIPT
phenanthroline based copper complexes [29]. This redox couple allowed to double the efficiency of
DSSCs with respect to [Cu(2,9-dimethyl-1,10-phenantholine)2]+/2+ in combination with a π-
extended benzothiadiazole dye [30] or a ruthenium dye working in the presence of an iron(II)
comediator [29]. Also, coupling zinc porphyrin dyes with the related [Cu(2-mesityl-1,10-
phenantholine)2]+/2+ redox couple afforded efficient DSSCs, exceeding the performance reached by
-
Co2+/3+ and I-/I3 reference electrolytes [31].
These observations prompted us to undertake a deeper investigation of the properties of a series of
homoleptic copper complexes based on 1,10-phenanthrolines, where the sole 2 position was
substituted by various aryl and alkyl moieties. In particular, the electrochemical behavior of copper
complexes substituted with mesityl, tolyl, phenyl or n-butyl groups (1/2, 3/4, 5/6, 7/8, Chart 1) was
analyzed and related to the photoelectrochemical characteristics of DSSCs sensitized by π-extended
benzothiadiazole dye (G3 or G4, Chart 1) [32,33].
(Chart 1)
2. Experimental
2.1 Synthesis of copper complexes redox mediators (1-8)
2.1.1 General comments
The hindered 2-substituted phenanthroline ligands were prepared by nucleophilic aromatic
substitution of the commercially available 1,10-phenanthroline with the suitable lithium derivative.
Mesityllithium and tolyllithium were prepared by transmetallation reaction between their bromo-
derivative and tert-butyllithium in diethylether (Et2O) as previous described [29] whereas copper
complexes 1 and 2, with 2-mesityl-1,10-phenanthroline as ligand, were synthesized as reported
[31].
Solvents for the syntheses were dried by standard procedures: N,N-dimethylformamide (DMF) was
dried over activated molecular sieves and toluene was distilled over Na/benzophenone. All reagents
5

ACCEPTED MANUSCRIPT
were purchased from Sigma-Aldrich and were used without further purification. Reactions requiring
anhydrous conditions were performed under argon or nitrogen. Thin layer chromatography (TLC)
was carried out with pre-coated Merck F254 silica gel plates. Flash chromatography (FC) was
carried out with Macherey-Nagel silica gel 60 (230-400 mesh).
2.1.2 General procedure for Cu(I) complexes
The precursor CuCl was dissolved under argon in acetonitrile and added dropwise to a stirred
solution of the suitable 2-substituted-1,10-phenanthroline in dry acetonitrile. The red solution was
stirred at room temperature for ca. 40 minutes. The solvent was eliminated under reduced pressure
and the solid was dissolved in the minimum amount of a mixture H2O/EtOH. After addition of an
aqueous solution of NaPF6, a red precipitate was formed and, after 2 hours, filtered off, washed with
H2O and Et2O and finally dried under vacuum.
2.1.2.1 [Cu(2-tolyl-1,10-phenanthroline)2]PF6 (3)
It was obtained starting from CuCl (12.4 mg, 0.126 mmol) in 1.8 mL of acetonitrile and 2-tolyl-
1,10-phenanthroline (69.8 mg, 0.258 mmol) in 9.5 mL of acetonitrile. The product was dissolved in
a mixture H2O/EtOH 1/1 (v/v) (ca. 10 mL) and NaPF6 (63mg in 2mL of water) was added. The
pure product was obtained as a dark red powder in 70% yield.
1H-NMR (400 MHz, CD2Cl2) δ 8.93 (d, J =4.39 Hz, 2H), 8.55-8.51 (m, 4H) 8.10(d, J =8.78 Hz,
2H), 8.01 (d, J =8.78 Hz, 2H), 7.88 (dd, J =8.10, 4.3 Hz, 2H), 7.74 (d, J =8.16 Hz, 2H), 6.56 (d, J
=7.40 Hz, 2H), 6.42 (t, J =7.39 Hz, 2H), 6.35 (d, J =7.49 Hz, 2H), 6.06 (t, J =7.39 Hz, 2H), 1.85(s,
6H). 13C-NMR (100 MHz, CD2Cl2) δ 157.92, 148.63, 143.61, 143.07, 139.70, 136.58, 134.67,
129.42, 128.7, 127.76, 127.64, 126.56, 126.62, 125.14, 123.90, 19.98. Elemental analysis: Calcd.
(%) for C38H28CuF6N4P : C 60.92; H 3.77; N 7.48; found: C 61.01; H, 3.77; N 7.47. MS (FAB+)
m/z: 603.13 Found: 603.
6

ACCEPTED MANUSCRIPT
2.1.2.2 [Cu(2-phenyl-1,10-phenanthroline)2]PF6 (5)
It was obtained starting from CuCl (19.9 mg, 0.201 mmol) in 2.7 mL of acetonitrile and 2-phenyl-
1,10-phenanthroline (106 mg, 0.413 mmol) in 15. mL of acetonitrile. The product was dissolved in
a mixture H2O/EtOH 1/1 (v/v) (ca. 15 mL) and NaPF6 (101 mg in 4 mL of water) was added. The
pure product was obtained as a dark red powder in 88% yield.
1H-NMR (400 MHz, CD2Cl2) δ 9.06 (dd, J =4.3 Hz, J= 1.4 Hz 2H), 8.65 (d, J =8.45 Hz, 2H), 8.47
(d, J =8.45 Hz, 2H), 8.09 (d, J =8.08 Hz, 2H), 8.04 (d, J =8.08 Hz, 2H), 7.96 (dd, J =8.0, 1.4 Hz,
2H), 7.85 (d, J =8.08 Hz, 2H), 7.23 (d, J =7.4 Hz, 4H), 6.66 (t, J =7.4 Hz, 2H), 6.38 (t, J =7.4 Hz,
4H). 13C-NMR (400 MHz; CD3CN): δ 158.20, 148.83, 143.71, 143.13, 138.68, 136.64, 134.77,
129.40, 128.55, 127.76, 127.68, 126.59, 125.66, 125.14, 124.01. Elemental analysis: Calcd. (%) for
C36H24CuF6N4P (721.11): C, 59.96; H, 3.35; N, 7.77; found: C, 60.03; H, 3.34; N, 7.80. MS (ESI+)
m/z: 575.13 Found: 575.49.
2.1.2.3 [Cu(2-n-butyl-1,10-phenanthroline)2]PF6 (7)
It was obtained starting from CuCl (70 mg, 0.366 mmol) in 8 mL of acetonitrile and 2-n-butyl-
1,10-phenanthroline (300 mg, 1.27 mmol) in 13.3 mL of dry acetonitrile. The solid was dissolved in
a mixture H2O/EtOH 1/2 (v/v) (ca. 50 mL) and NaPF6 (305 mg in 9 mL of water) was added. The
pure product was obtained as a dark orange powder in 87% yield.
1H-NMR (400 MHz; MeOD): δ 9.06 (d, J=4.1 Hz, 2H), 8.79 (d, J=7.8 Hz, 2H), 8.69 (d, J=8.3 Hz,
2H), 8.23 (d, J=9.0 Hz, 2H), 8.20 (d, J=9.0 Hz, 2H), 8.00 (dd, J1=7.8 Hz, J2=4.1 Hz, 2H), 7.93 (d,
J=8.3 Hz, 2H), 2.78-2.64 (m, 4H), 1.48-1.22 (m, 4H), 0.69-0.48 (m, 4H), 0.22 (t, J=7.4 Hz, 6H).
Elemental Analysis: Calcd. for C32H32CuF6N4P: C 56.43, H 4.74, N 8.40. Found. C 56.55, H 4.75,
N 8.23. MS (ESI+) m/z: 535.19 Found: 535.57.
2.1.3 General procedure for Cu(II) complexes
7

ACCEPTED MANUSCRIPT
The phenanthroline ligand was dissolved in acetonitrile and a solution of CuSO4*5H2O in water
was slowly added. The light green mixture was stirred at room temperature for 1h and then it was
concentrated at reduced pressure at about 2 mL and a water solution of NaPF6 was added. The
resulting solution was stirred for 2h and then the green precipitate was filtered, washed successively
with H2O and Et2O and finally dried under reduced pressure.
2.1.3.1 [Cu(2-tolyl-1,10-phenanthroline)2](PF6)2 (4)
Obtained starting from 2-tolyl-1,10-phenanthroline (31.0 mg, 0.114 mmol) in acetonitrile (9 mL)
and CuSO4*5H2O (14.2 mg, 0.057 mmol) in 2 mL of water. After addition of NaPF6 (40 mg in 1
mL of water), the product was obtained as a green solid in 40% yield.
Elemental Analysis: Calcd. for C38H28CuF12N4P2: C 51.04, H 3.16, N 6.27. Found. C 51.25, H 3.17,
N 6.30. MS (FAB+) m/z: 603.
2.1.3.2 [Cu(2-phenyl-1,10-phenanthroline)2](PF6)2 (6)
Obtained starting from 2-phenyl-1,10-phenanthroline (30 mg, 0.111 mmol) in acetonitrile (8 mL)
and CuSO4*5H2O (17.5 mg, 0.07 mmol) in 1.5 mL of water. After addition of NaPF6 (56 mg in 1
mL of water), the product was obtained as a green solid in 68% yield.
Elemental Analysis: Calcd. for C36H24CuF12N4P2: C 49.92, H 2.79, N 6.47. Found. C 49.97, H 2.81,
N 6.46. MS (ESI+) m/z: 575.42.
2.1.3.3 [Cu(2-n-butyl-1,10-phenanthroline)2](PF6)2 (8)
Obtained starting from 2-n-butyl-1,10-phenanthroline (38.0 mg, 0.161 mmol) in acetonitrile (3 mL)
and CuSO4*5H2O (20 mg, 0.08 mmol) in 2 mL of water. After addition of NaPF6 (58 mg in 1 mL
of water), the product was obtained as a green solid in 70% yield.
Elemental Analysis: Calcd. for C32H32CuF12N4P2: C 46.53, H 3.90, N 6.78. Found. C 46.61, H 3.92,
N 6.80. MS (ESI+) m/z: 535.45.
8

ACCEPTED MANUSCRIPT
2.2 Electrochemical characterization
Cyclic voltammetry, CV, and electrochemical impedance spectroscopy, EIS, for the electrochemical
characterization of copper complexes were performed in a customized three-electrode minicell
filled with 2-3 mL of a solution of acetonitrile as solvent with 0.1 M tetrabutylammonium
hexafluorophosphate, [TBA]PF6 as supporting electrolyte. The working solution was well deareated
by bubbling nitrogen before starting each measurement. A nitrogen blanket was maintained over the
solution during the scans as well. Teflon-embedded glassy carbon electrode GC (geometric surface
area 0.071 cm2) were used as working electrode in combination with a platinum wire as counter
electrode and a saturated calomel electrode, SCE, as reference one. To avoid water and chloride
leakage into the working solution, SCE was inserted into a glass jacket ending with a porous frit
and filled with the same blank electrolytic solution (i.e. solvent and supporting electrolyte). The
recorded potentials were subsequently referred to the intersolvental reference redox couple Fc+|Fc
(ferricenium|ferrocene) added as external standard (ca. 1ꢀ10−3 M) at the end of each daily measure.
The measures were performed with a potentiostat/galvanostat PGSTAT302N (Metrohm Autolab,
The Netherlands) controlled by a PC through GPES and NOVA 1.10 software, equipped with a
frequency response analyzer module (FRA2) for EIS measurements. Staircase CV were performed
at different sweep potentials (0.02-2 V s–1) with a potential step of 0.001-0.002 V. An instrumental
compensation of the resistance (i.e. positive feedback) was carefully performed in order to
minimize the ohmic drop between the working and reference electrode. EIS spectra were recorded
superimposing to a continuous potential bias set at the half-wave potential, E1/2, of the analyzed
compound a sinusoidal alternating potential signal of 0.01 V amplitude and frequency ranging from
1ꢀ105 to 1ꢀ10–1 Hz. Sixty logarithmically distributed single sine frequencies were employed for
recording each spectrum. NOVA 1.10 software was used to validate EIS data (through Kronig-
Kramers analysis) and to fit them with a Randles-type equivalent circuit to estimate the charge
transfer resistance, Rct, at the electrolyte/electrode interface.
9

ACCEPTED MANUSCRIPT
2.3 Dye solar cells fabrication and photoelectrochemical characterization
FTO-coated glass substrates (15 ꢁ sq–1, purchased from XIN YAN TECHNOLOGY LTD) were
cleaned in a detergent solution for 15 min and then in ethanol for 30 min using an ultrasonic bath.
Mesoporous TiO2 electrodes were prepared by doctor blading a commercial paste, namely Dyesol
30NRD up, and subjecting them to thermal sintering at 450 °C for 40 min. The average thickness of
the electrodes used in the reported batch of devices was 12 ꢂm. After this first thermal sintering, all
the samples were treated with a 40 mM aqueous solution of TiCl4 for 30 min at 70° C, rinsed again
with water and ethanol and then subjected to a second sintering process at 500°C for 30 min. After
cooling the mesoporous electrodes were immersed into the dye solution (G3 or G4, 0.2 mM) along
with chenodeoxycholic acid (30 or 10 mM, respectively) in tetrahydrofuran for 8 h. The counter
electrodes were prepared by sputtering a 50 nm Pt layer on a hole-drilled cleaned FTO plate. The
two electrodes were then assembled together in a sandwich configuration and sealed upon heating a
thermoplastic gasket made of a hot-melt ionomer-class resin (Surlyn 50-ꢂm thickness). The
electrolytic solution was injected through the hole on the counter electrode glass. For symmetrical
cells, two identical Pt-coated FTO plates were assembled, using a 100-ꢂm thick gasket.
Photocurrent versus voltage, j-E, measurements were performed using a Keithley unit (Model 2400
Source Meter), while a Newport Sol3A Class AAA Solar Simulator (Model 94063A equipped with
a 1000W xenon arc lamp) served as a light source, scanning potential from open-cirucit to short
circuit conditions. The light intensity (or radiant power) was calibrated to 100 mW cm−2 using as
reference a Silicon standard solar cell. Electrochemical analyses (on both DSSCs and symmetrical
cells) were performed using an Autolab PGSTAT 302N, and digitally processed using the NOVA
1.10 software. EIS spectra were recorded by sweeping the frequency from 50 kHz to 0.01 Hz with a
logarithmic distribution, using a 10 mV amplitude sine wave potential modulation superimposed to
the bias potentials. EIS data were fitted to the equivalent circuit using the ZWiew® software.
10

ACCEPTED MANUSCRIPT
2.4 Transient absorption spectroscopy measurements
Transient absorption spectroscopy, TAS, was carried out by a previously described apparatus [34]
using the 532 nm harmonic of a nanosecond Nd:YAG laser (Continuum Surelite II). A typical pulse
energy of 2 mJ cm−2 was used. A 532 nm notch filter prevented laser light from reaching the
photomultiplier. Formation of the photo-oxidized dye was probed at 770 nm, by using the
monochromatic output of an Applied Photophysics f: 3.4 monochromator coupled to a pulsed Xe
arc lamp. Typically 30 laser shots, at a frequency of 0.2 Hz, were averaged to reach a good S/N
ratio.
Photoanodes were prepared by a standard procedure. FTO glass plates were washed in 2-propanol
for 10 min using an ultrasonic bath, then were heated at 450°C for 20 minutes. A compact TiO2
blocking underlayer was prepared by spin-coating a 0,3 M titanium(IV) isopropoxide solution in 1-
butanol (1000 rpm for 10 s, 2000 rpm for 20 s). Then the substrates were heated up at 500 °C for 15
min. The photoelectrodes were prepared by doctor blading, using a transparent TiO2 paste (Dyesol
18NR-T). The coated films were heated to 500 °C for 10 min with programmed temperature
ramping. After sintering, the electrodes were treated with 0.4 M TiCl4 overnight and heated again at
450° for 30 minutes. Once cooled, the electrodes were dipped for 15 minutes in a 0.2 mM solution
of the dye (G3) in tetrahydrofuran, containing chenodeoxycholic acid 30 mM.
TAS measurements on transparent thin films in the presence of the electron donor species in
acetonitrile were performed by drawing the electrolyte by capillarity inside the chamber (ca. 6−8
ꢂm), constituted by a glass slide pressed against the TiO2 photoanode. Thus, the negligible optical
path allowed to avoid interferences originated by the mediators’ absorbance.
Regeneration efficiency was calculated by applying the formula [35]:
kr'eg
kr'eg + krec
η
=
(1)
11

ACCEPTED MANUSCRIPT
where the pseudo first order recombination, krec, and regeneration, k’reg, rate constants were
obtained by fitting the experimental decay curves with a triexponential function (see also section
3.3 for further details):
y = A exp(−x /t1) + A2 exp(−x /t2 ) + A3 exp(−x /t3 ) + y0
1
(2)
in order to optimize S/N ratio and by individuating on the fitting curves the point corresponding to
half of the initial normalized amplitude (ꢃA/2) in whose correspondence τ1/2 is obtained. By
applying the following equation pseudo-first order rate constants are calculated as follows:
−1
k = (τ1/ 2
)
(3)
(4)
(5)
−1
krec = (τ1/ 2
)
blank
kr'eg = (k − krec )
The following solutions were tested :
0.1 M LiOTf in acetonitrile (blank); 0.15 M 1, 0.1 M LiOTf , in acetonitrile (1): 0.15 M 3, 0.1 M
LiOTf , in acetonitrile (3); 0.15 M 5, 0.1 M LiOTf , in acetonitrile (5); 0.15 M 7, 0.1 M LiOTf , in
acetonitrile (7); 0.15 M 1,3-dimethylimidazolium iodide, 0.1 M LiI, in acetonitrile (I–); 0.15 M
[Co(bpy)3](OTf)2, 0.1 M LiOTf , in acetonitrile (Co(II)).
3. Results and discussion
3.1 Optical spectroscopy
Optical features of complexes were studied by visible-NIR spectrophotometry. All copper(I)
complexes appeared from light to dark red solids; coherently the absorption spectrum recorded in
acetonitrile (Table 1) showed a broad band centred around 450 nm of intensity ca. 4.5ꢀ103 M–1 cm–1
attributed to a metal-to-ligand charge transfer transition (MLCT) [36]. The blue shift of complex 5,
contrary to expectation due to absence of any methyl group on the phenyl, points to a stronger
hindrance [37] imposed by the aromatic substituents than in 1 and 3 analogues as a result, probably,
of the free-rotation of the phenyl rings around the interanular bond (as confirmed by only three -Ph
signals in 1H-NMR) precluded to both tolyl and mesityl rings. The much higher molar absorption
12

ACCEPTED MANUSCRIPT
coefficient for the unique alkyl substituted compound 7 resembles the trend observed for some 2,9-
disubsituted analogues [38]. Interestingly, in 1 the absence of methyl groups in the 4,7 positions of
phenanthroline determines a bathochromic shift of the absorption respect to the related compound
[Cu(2-mesityl-4,7-dimethyl-1,10-phenantholine)2]+. Moreover, the asymmetric substitution of the
positions 2 and 9, in α of the coordinated nitrogen atoms, resulted in a less intense absorption if
compared to the di-substituted benchmark compound [Cu(2,9-dimethyl-1,10-phenanthroline)2]+,
reducing the competitive absorption of light by the mediators respect to dyes. This observation
further strengthens the assertion of a strict correlation between spectral intensity and symmetry of
the complex, and hence the distortion from the tetrahedral geometry induced by ligand substituents
[38]. Furthermore it suggests that ligand asymmetry decreases the intensity of MLCT bands
regardless of whether it is introduced in 4,7 [38] or 2,9 positions.
Oxidized complexes invariably showed a broad band at around 690-750 nm attributed to a ligand
field d-d transition (allowed by d9 configuration of Cu atom) significantly less intense than the
MLCT of copper(I) analogues (Table 1), eventually followed by a second weaker band at
wavelengths around 900-1000 nm. These spectra are consistent with previous published analogues
[29-31] and can be tentatively attributed to a five-coordinate geometry in solution, with acetonitrile
as ancillary ligand of a trigonal bipyramide [39,40].
(Table 1)
3.2 Electrochemistry
Electrochemical behaviour of the copper-based coordination compounds were studied through
cyclic voltammetry in acetonitrile. Concerning the anodic window (Figure 2), all complexes
presented a single wave attributed to an electron transfer centred on the metal core, in agreement
with literature [41], directly involved in both dye regeneration and cathodic reaction in DSSC
operation. The process is perfectly chemically reversible and the superimposition of the pattern of a
copper(I) compound with its oxidized counterpart is a clear proof of the mutual interconversion
13

ACCEPTED MANUSCRIPT
between the two species. A peculiar situation occurred only for the copper(II) compound 2, whose
electron transfer mechanism seems to be more complicated, directly involving the solvent as
already reported [31].
(Figure 2)
The oxidation potential of copper complexes (i.e. half-wave potential, E1/2) is modulated by the
nature of the 2-substituted-1,10-phenanthrolines through a combination of electronic and steric
effects [41]. With respect to E1/2 = –0.02 V vs Fc+|Fc of [Cu(2-mesityl-4,7-dimethyl-1,10-
phenantholine)2]+ [29,30,41], complexes 1-8 exhibit positively shifted half-wave potentials ranging
from 0.06 to 0.11 V vs Fc+|Fc (Table 2). Such anodic shift of the metal centered Cu(I)/(II) process
can be reasonably attributed to the absence of electron rich methyl groups in 4,7 positions which
increases the electron withdrawing properties of the aromatic ligands via dπ-π* bonding
interactions. In a similar way, the progressive reduction of the number of –CH3 on the 2-aryl moiety
on going from 1 to 3 to 5 can be, at least partially, the reason for the E1/2 positive shift. Nevertheless
it is reasonable to take into account also the different steric hindrance and conformational rigidity of
the substituents in position 2, which affect the geometry modification around the copper center,
favouring or disfavouring a specific oxidation state of the copper atom and hence modulating both
thermodynamic and kinetics of electron transfer. In fact, an important feature of d10 copper(I)
complexes is a significant nuclear rearrangement due to the change of coordination geometry, from
a tetrahedral to a tetragonal configuration upon oxidation to give d9 copper(II) complexes [10].
However, steric constraints, such as the presence of sterically challenging groups in the α positions
immediately adjacent to the coordinating nitrogen atoms of bipyridines or phenanthrolines, is an
elegant way to control coordination geometry and its changes going from one oxidation state of the
copper to the other [10]. In a similar way, the modulation of the inner-sphere reorganization energy
of the complexes by substituents in position 2 regulates the rate of the electron transfer, according
to the Marcus theory [42,43]. Table 2 summarizes the electron transfer rate constants, kheter
,
14

ACCEPTED MANUSCRIPT
obtained by Electrochemical Impedance Spectroscopy, EIS, measurements in acetonitrile on glassy
RT
F2CRct A
carbon electrode. The heterogeneous rate constant was calculated according to kheter
=
,
where R is the universal gas constant, T is the absolute temperature, F is the Faraday constant, C is
the copper(I) complex concentration expressed in mol cm–3, Rct the charge transfer resistance at an
applied bias equal to E1/2 obtained by fitting EIS spectrum with the classical Randles circuit, and A
the geometric area of the electrode in cm–2. The fastest electron transfer occurred with 5, kheter
=1.9ꢀ10–2 cm s–1 at E1/2, comparable with the heterogeneous kinetics observed with the benchmark
[Cu(2,9-dimethyl-1,10-phenanthroline)2]+ complex, while compound 7 with n-butyl chains on the
ligands resulted to be the most sluggish of the series, with a three-times slower kinetics. A
quantitative explanation that links structure of copper complexes, half-wave potential and electron
transfer constant rate is out of the aim of the present work and will be carried out on a much larger
series of complexes, where structure-activity relationships could be established with a better degree
of certainty with respect to the small series herein presented. Nevertheless it is possible to note that
the simple assertion relating higher half-wave potential to faster kinetic constant is not applicable,
as exemplified by complex 7. As observable from the relatively high value of E1/2 of complex 7,
combined with its low kheter, the alkyl substituents induce a geometry that well stabilizes the Cu(I)
form but, at the same time, their flexibility (respect to the rigid planarity of aryl rings) does not
efficiently hamper the geometry modification from Cu(I) to Cu(II) and vice versa, increasing the
inner sphere reorganization energy, which ultimately turns into an higher activation barrier for the
electron transfer. Among complexes 1, 3, 5 with 2-aryl substituted-1,10-phenanthrolines the rate
constant increases with decreasing the number of methyl groups on the phenyl ring and so with
reducing the steric hindrance of the substituents in position 2 while increasing their rotational
freedom. For sake of comparison the heterogeneous rate constant of a cobalt(II) complex,
[Co(bpy)3]2+ (bpy = 2,2’-bipyridine), was evaluated with the same protocol. Its value, kheter=8ꢀ10–3
cm s–1, is comparable with that of the two slowest copper(I) complexes, 1 and 7.
15

ACCEPTED MANUSCRIPT
(Table 2)
3.3 Dye regeneration
In a DSSC the reduced form of the mediator couple is tasked to react with the oxidized sensitizer,
S+, chemisorbed on the TiO2 surface, in order to regenerate the starting form of the dye, S, as
schematically depicted by the process (3) in Figure 1. Starting from some promising results [30, 32,
33] the fully-organic dye G3 [32] was selected as a fruitful candidate to test the electron transfer
properties of this new series. The capability of copper mediators to regenerate the sensitizer was
evaluated by transient absorption spectroscopy (TAS), in comparison with reference donor species
like [Co(bpy)3]2+ and I–. In general, the oxidized dye recovery kinetics are described by a
multiexponential behavior (like equation 2) essentially determined by the surface heterogeneity of
the mesoscopic TiO2 film, which results in a distribution of surface states from which
recombination occurs with different rate constants. Due to this fact, recombination kinetics on TiO2
have been described with the Kohlrausch William Watts (KWW) function (stretched exponential)
by some authors [44]. In the absence of electron donors (blank electrolyte), oxidized dye recovery
occurs only by recombination with the photoinjected electron (krec, equation 4) according to
G3+ + e– TiO2 ꢀ G3
as schematically depicted by (6) in Figure 1.
Dye recovery kinetics monitored at 770 nm in a blank solution (i.e. acetonitrile with LiOTf)
revealed a τ1/2,blank around 56 ꢂs (equation 4). Addition of electron donor species (1, 3, 5, 7,
[Co(bpy)3]2+, I–) invariably reduces the lifetime, consistently with the electron transfer (k’reg,
equation 5) from the donor (D) to the oxidized dye (Figure 3) according to the competitive process
depicted by (5) in Figure 1
G3+ + D ꢀ G3 + D+
Copper-based electron donors are characterized by a lower driving force for dye regeneration with
respect to the two reference systems, namely cobalt (II) and iodide, being their oxidation half-wave
16

ACCEPTED MANUSCRIPT
potential at least 120 mV and 340 mV more positive than [Co(bpy)3]2+ and I–, respectively (using
–
E
1/2 I–/I3 = –0.28 V vs Fc+|Fc, converted from 0.35 V vs NHE [33],[45]). Nevertheless, the kinetic
competition between dye regeneration and recombination is in all cases favourable to a nearly
quantitative regeneration efficiency (Table 3), expressed according to equation 2. All copper(I)
complexes are able to assure regeneration of 50 % of the initial G3+ population chemisorbed on a
transparent TiO2 photoanode on a time interval varying between 0.7-2 ꢂs, i.e. more than 20 times
faster than the interfacial recombination. This implies regeneration efficiencies higher than 96% for
the copper series, comparable or even higher than those obtained with cobalt (II) and I– reference
solutions (Table 3, where η ranges from 99% for 5 to 92% for I–). Thus the differences in η,
measured in the absence of applied forward bias, are always rather minor and should not cause
significant variations in cell performance as long as dye regeneration is concerned. According to the
semiclassical Marcus theory, the rate of electron transfer in a bimolecular reaction is given by the
product of the electronic coupling, HAB, proportional to the overlap between the wavefunctions of
the electron donor and of the acceptor, with the nuclear activation term, where the Gibbs free
(λ
+ ∆G0 )2
energy of activation, ꢃG#, defined as ∆G# =
appears. Within the copper series, both
4
λ
HAB and the standard Gibbs free energy, ꢃG0, for each dye/copper donor are expected to be
essentially constant (being E1/2 of copper complexes very similar especially for 3, 5, 7) while slight
differences in the reorganization energy, λ, bearing the dominant contribution of the nuclear
rearrangement that follows the Cu(I) ꢀCu(II) transition, probably contribute to explain the
differences in the regeneration rate constant. Other more subtle factors, like local surface
interactions between the mediator with the sensitized oxide surface and differences in the
permeation of the titania mesopores by the liquid electrolyte may influence the outcome of
regeneration rate constants.
Interestingly, it appears that the best pseudo first-order kinetic constant, k’reg, is observed with
complex 5 which also displays the highest heterogeneous electron transfer constant, kheter, on glassy
17

ACCEPTED MANUSCRIPT
carbon electrode. However developing a systematic correlation between the charge transfer kinetics
observed at conventional working electrodes, like glassy carbon, and the regeneration rate observed
at sensitized interfaces is difficult, and may lead to oversimplifications. Compared to glassy carbon,
the polar surface of TiO2 presents entirely different properties and the presence of a chemisorbed
dye further adds to the interfacial complexity. Specific differences in the double layer structure
between the two different electrodic materials and the specific electronic coupling which is
established between the oxidized G3 dye and the redox mediator are not taken into account in the
voltammetric experiment at GC electrode. For example, in the case of the quasi-reversible cobalt
(II)/(III) complexes, a large influence of the double layer structure on the heterogeneous charge
transfer involving cobalt mediators has been reported [46] while modification of otherwise
electrochemically inert substrates with certain molecular species was observed to promote a
significant electrocatalysis towards the Co(II)/(III) couple [47,48]. A dependence of the
electrochemical response on the nature of the electrodic material was also observed with both cobalt
complexes [49] and some Cu(I)/(II) redox couples [29,30], suggesting also for the present case
electrode-specific activation barriers for the heterogeneous electron transfer.
(Figure 3)
(Table 3)
3.4 Cu-complexes as electron mediators in DSSCs
Different electrolytes were prepared with each of the copper redox couples, dissolving 0.15 M of
the copper(I) species (1, 3, 5, 7) and 0.015 M of the corresponding copper(II) form (2, 4, 6, 8) in
acetonitrile. Essentially, 0.15 M concentration is close to the solubility limit for most of the copper
complexes at RT in the solvent of our choice. Lithium trifluoromethanesulphonate (LiOTf) 0.1 M
and tert-butylpyridine (tbp) 0.25 M were also added as additives in each formulation. For sake of
–
comparison, reference electrolytes based on cobalt, [Co(bpy)3]2+/3+-El, and iodine/triiodine, I–/I3 -El,
18

ACCEPTED MANUSCRIPT
were not optimized but prepared in identical way, i.e. equimolar concentration, with the same
additives. This avoids spurious effects due to the mismatch in electrolyte composition and mediator
concentration allowing to highlight the intrinsic impact of the properties of the various redox
–
electrolytes on cell efficiency. Along with dye G3 (PCE up to 8.6% with an optimized I–/I3
electrolyte, under 100 mW cm−2 simulated AM 1.5G irradiation [32]) electrolytes were tested in
solar devices sensitized with its analogue, G4 (PCE up to 9.5% [33]), bringing one hexyl chain on
the π–conjugated central scaffold with the aim to further hamper charge recombination by the
oxidized mediator (Chart 1).
jE curves and related efficiency parameters for G3-sensitized DSSCs are reported in Figure 4 and
Table 4, respectively.
(Figure 4)
(Table 4)
Impedance analysis performed at Voc in the dark on G3-sensitized devices shows that differences in
kheter of copper complexes observed on the ideally inert GC electrode are partially equilibrated in
DSSC, probably due to the quite different nature of the interfaces on which the electron transfer
occurs, as previously discussed in section 3.3. In fact charge recombination (process (5) in Figure 1)
for Cu-based mediators (i.e. the process occurring at medium frequencies, Figure 4) is characterized
by comparable time constant of about 10 ms (equal to the inverse of the angular frequency at –
Z’’max, Figure 4), similar to that of the control couple [Co(bpy)3]2+/3+. As a consequence
photovoltages well correlate with the E1/2 of the couples, i.e. the more positive is E1/2 the higher is
Voc.
Interestingly most Cu-based redox couples are very efficient, outperforming G3-sensitized control
–
cells filled with both I–/I3 and [Co(bpy)3]2+/3+ reference electrolytes (prepared with comparable
19

ACCEPTED MANUSCRIPT
–
concentrations). In particular, the slower charge recombination of I–/I3 -El (Figure 4) does not
guarantee the highest open circuit potential, being negatively affected by the too negative E1/2 of the
couple with respect to copper complexes. The best performing redox couple 3/4 reached a 6.0%
power conversion efficiency with both dyes, slightly improving the performance of our previous
record couple 1/2 [31], mainly thanks to a 0.03-0.04 V improvement in Voc. These two couples
show similar electron transfer kinetics (Table 2) and comparable charge transfer recombination
resistances (Figure 4) in DSSCs. This suggests that the improved performance of 3/4–El with
respect to 1/2-El is related to lower overvoltage losses in the dye regeneration stage (due to its more
positive oxidation potential), allowing to extract higher photovoltage while maintaining a nearly
quantitative dye regeneration efficiency.
An anomaly in the G3-sensitized cells is 5/6-El which is characterized by a quite low short circuit
current density, which can be tentatively related to a poorer electron injection potentially
attributable to an upshift of the TiO2 conduction band edge energy. Such hypothesis would correlate
well with the smaller drop in jsc recorded with G4 (Table 4), characterized by a ca. 0.1 eV higher
LUMO energy than G3 [33], which anyhow assures a good driving force for electron injection into
the semiconductor. Another reason at the basis of the low photocurrent recorded with 5/6-El may
reside in a larger diffusional resistance of such redox couple, revealed in symmetric Pt/Pt cell
(Figure 5).
(Figure 5)
In general, the Fill Factor of the G3 sensitized cells varies between 59% and 69%. The highest
value (69%) is observed with 5/6-El, which however produces the lowest current. The cell Fill
Factor is a measurement of the cell quality in terms of overvoltage losses in the cell, which can be
expressed in the form of jR drops. The smaller these drops are, the better is FF. In particular, R is
the sum of all the serial resistances of the cell, including ohmic, charge transfer and mass transport
20

ACCEPTED MANUSCRIPT
resistances. It is not surprising that the best Fill Factor is thus observed in the presence 5/6-El,
endowed with a ca. 30% lower photocurrent with respect to the best couples, since the jR drop
scales with the photocurrent. In our case, the best compromise is found with 3/4-El and 7/8-El for
which FF is still 62 and 66% respectively, while the photocurrent is at least 20% higher than 5/6-El.
–
Contrary to what previously reported for optimized I–/I3 -El [33], the introduction of hexyl chain
along the π-conjugated backbone of the G4 dye does not bring the expected positive effects on the
recombination by Cu-based electrolytes. In fact PCEs remain equal (3/4-El) or even slightly lower
(in all the other cases) than with the simpler G3 sensitizer (Table 4), essentially due to a lower
photovoltage. In the case of 5/6-El the increased jsc is offset by a decreased photovoltage and FF,
and the resulting PCE is essentially the same, but in the case of 7/8-El the drop in photovoltage
(0.08 V) and slightly reduced FF (63% for G4 vs 66% for G3) pairs with an unchanged
photocurrent density (10.1 mA cm–2). The lower Voc generally observed with G4 could be
reasonably explained by a faster charge recombination involving copper(II) complexes on the
titania surface.
4. Conclusions
Four Cu+/Cu2+ redox couples constituted by homoleptic complexes of 2-substituted-1,10-
phenanthrolines were synthesized. Functionalization of the positions adjacent to nitrogen atoms
allows to modulate both optical and electrochemical features of the related copper complexes. As a
result, combination of visible absorption spectroscopy and cyclic voltammetry shows that the
asymmetric substitution of ligands (i.e. functionalization of internal position 2) is an elegant way to
produce copper-based redox couples with valuable features as electron mediators in dye-sensitized
solar cells. In particular 2-subsituted ligands decrease intensity of the MLCT bands of related Cu(I)
coordination compounds, resembling the already known effect of the asymmetry induced by the 4,7
external positions of phenanthrolines. This effect is amplified if aryl groups are preferred to alkyl
ones. The electrochemical study reveals that E1/2 and kheter are not directly proportional, but they are
21

ACCEPTED MANUSCRIPT
influenced by the geometry of the complex and by its flexibility, which are modulated not only by
the steric hindrance but also by the rotational freedom of the internal substituents.
As a result 2-aryl derivatives effectively combine optimal optical and electrochemical properties,
thanks to lower molar extinction coefficients, sufficiently positive oxidation potentials and well-
balanced electron transfer kinetics resulting in nearly quantitative dye regeneration efficiencies
under short circuit conditions. While the details of the mechanistic correlation between synthetic
design, basic electrochemical properties, and the resulting power conversion in the solar devices
remained elusive, due to the presence of different complex interfaces and the interlacing of several
processes involved in cell functioning, it can be affirmed that the proposed synthetic design was
generally successful in producing mediators endowed with a reduced overvoltage loss associated to
the dye regeneration process. As a result 3/4-El neatly shows a 20% and 15% increase in PCE
−
respect to the common [Co(bpy)3]2+/3+ and I−/I3 couples, respectively, formulated at comparable
concentration, overcoming also our previous best performing copper-based mediator, 1/2-El, thanks
to a more positive half-wave potential.
The extension of these studies to a wider Cu(I)/(II) series, the back up by DFT calculations, as well
as the examination of the dye regeneration kinetics in the working device under applied bias are
expected to bring a clearer correlation between the basic optical and electrochemical properties of
these species with cell performance, paving the way towards a further generation of optimized
copper-based mediators.
Acknowledgements
A. C. thanks Università degli Studi di Milano (Piano Sostegno alla Ricerca 2015–17-LINEA 2
Azione A – Giovani Ricercatori) for financial support. R.G. and G.P.S. acknowledge funding from
the Austrian Research Promotion Agency-FFG (Bridge-Early Stage COMPOSTRONICS cod.
5730587). The “Ministero degli Affari Esteri e della Cooperazione Internazionale” (bilateral project
22

ACCEPTED MANUSCRIPT
Italy-India, Prot. nr. MAE0104617) is also acknowledged for financial support. We thank both
Regione Lombardia and Fondazione Cariplo for the use of instrumentation purchased through the
SmartMatLab Centre project (2014). POR-FESR HP Solar and Solarchem are gratefully
acknowledged (UNIFE).
References
[1] Dye-sensitized Solar Cells, ed K. Kalyanasundaram, EPFL Press, Lausanne, 2010.
[2] Global Dye-sensitized Solar Cells (DSSC) Market Professional Survey 2016 Industry Trend and
Forecast 2021, QY Research (2016).
[3] A. Fakharuddin, R. Jose, T.M. Brown, F. Fabregat-Santiago, J. Bisquert, A perspective on the
production of dye-sensitized solar modules. Energy Environ. Sci., 7 (2014), pp. 3952-398.
[4] B. O’Regan, M. Grätzel, A low-cost, high-efficiency solar cell based on dye-sensitized colloidal
TiO2 films, Nature 353 (1991) 737–740.
[5] M. Grätzel, Recent Advances in Sensitized Mesoscopic Solar Cells. Accounts Chem. Res., 42
(2009) 1788-1798.
[6] G. C. Vougioukalakis, A. I. Philippopoulos, T. Stergiopoulos , P. Falaras, Contributions to the
development of ruthenium-based sensitizers for dye-sensitized solar cells, Coord. Chem. Rev. 255
(2011) 2602-2621.
[7] L.-L. Li and E. W.-G. Diau, Porphyrin-sensitized solar cells, Chem. Soc. Rev. 42 (2013) 291-
304.
[8] C. E. Housecroft, E. C. Constable, The emergence of copper(I)-based dye sensitized solar cells,
Chem. Soc. Rev. 44 (2015) 8386-8389.
[9] M. Sandroni, Y. Pellegrin and F. Odobel, Heteroleptic bis-diimine copper(I) complexes for
applications in solar energy conversion, C. R. Chimie 19 (2016) 79-93.
23

ACCEPTED MANUSCRIPT
[10] M. Magni, P. Biagini, A. Colombo, C. Dragonetti, D. Roberto, A. Valore, Versatile copper
complexes as a convenient springboard for both dyes and redox mediators in dye sensitized solar
cells, Coord. Chem. Rev. 322 (2016) 69–93.
[11] C.-Y. Chen, M. Wang, J.-Y. Li, N. Pootrakulchote, L. Alibabaei, C. Ngoc-le, J.-D. Decoppet,
H.-H. Tsa, C. Graetzel, C.-G. Wu, S. M. Zakeeruddin, M. Graezel, Highly Efficient Light-
Harvesting Ruthenium Sensitizer for Thin-Film Dye-Sensitized Solar Cells, ACS Nano 3 (2009)
3103-3109.
[12] Q. Yu, Y. Wang, Z. Yi, N. Zu, J. Zhang, M. Zhang and P. Wang, High-Efficiency Dye-
Sensitized Solar Cells: The Influence of Lithium Ions on Exciton Dissociation, Charge
Recombination, and Surface States, ACS Nano 4 (2010) 6032-6038.
[13] H. Ozawa, Y. Okuyama and H. Arakawa, Dependence of the efficiency improvement of black-
dye-based dye-sensitized solar cells on alkyl chain length of quaternary ammonium cations in
electrolyte solutions, ChemPhysChem, 15 (2014) 1201-1206.
[14] C.A. Bignozzi, R. Argazzi, R. Boaretto, E. Busatto, S. Carli, F. Ronconi, S. Caramori, The role
of transition metal complexes in dye sensitized solar devices, Coord. Chem. Rev. 257 (2013) 1472–
1492.
[15] B. Pashaei, H. Shahroosvand, P. Abbasi, Transition metal complex redox shuttles for dye-
sensitized solar cells, RSC Adv. 5 (2015) 94814–94848.
[16] T.W. Hamann, The end of iodide? Cobalt complex redox shuttles in DSSCs, Dalton Trans. 41
(2012) 3111-3115.
[17] L. Giribabu, R. Bolligarla, M. Panigrahi, Recent Advances of Cobalt(II/III) Redox Couples for
Dye-Sensitized Solar Cell Applications, Chem. Rec. 15 (2015) 760–788.
[18] A. Yella, H.-W. Lee, H. N. Tsao, C. Yi, A. K. Chandiran, Md. K. Nazeeruddin, E. W.-G. Diau,
C.-Y. Yeh, S. M. Zakeeruddin, M. Graetzel, Porphyrin-sensitized solar cells with cobalt (II/III)-
based redox electrolyte exceed 12 percent efficiency, Science 334 (2011) 629-634.
24

ACCEPTED MANUSCRIPT
[19] S. Mathew, A. Yella, P. Gao, R. Humphry-Baker, B. F. E. Curchod, N. Ashari-Astani, I.
Tavernelli, U. Rothlisberger, M. K. Nazeeruddin, M. Grätzel, Dye-sensitized solar cells with 13%
efficiency achieved through the molecular engineering of porphyrin sensitizers, Nature Chemistry,
6 (2014) 242-247.
[20] K. Kakiage, Y. Aoyama, T. Yano, K. Oya, J. Fujisawa, M. Hanaya, Highly-efficient dye-
sensitized solar cells with collaborative sensitization by silyl-anchor and carboxy-anchor dyes,
Chem. Commun. 51 (2015) 15894–15897.
[21] L. Kavan, Y.Saygili, M. Freitag, S.M. Zakeeruddin, A. Hagfeldt, M. Grätzel, Electrochemical
properties of Cu(II/I)-based redox mediators for dye-sensitized solar cells. Electrochim. Acta 227
(2017) 194-202.
[22] S. Hattori, Y. Wada, S. Yanagida, S. Fukuzumi, Blue Copper Model Complexes with
Distorted Tetragonal Geometry Acting as Effective Electron-Transfer Mediators in Dye-Sensitized
Solar Cells, J. Am. Chem. Soc. 127 (2005) 9648–9654.
[23] Y. Bai, Q. Yu, N. Cai, Y. Wang, M. Zhang, P. Wang, High-efficiency organic dye-sensitized
mesoscopic solar cells with a copper redox shuttle, Chem. Commun. 47 (2011) 4376–4378.
[24] M. Freitag, F. Giordano, W. Yang, M. Pazoki, Y. Hao, B. Zietz, M. Graetzel, A. Hagfeldt, G.
Boschloo, Copper Phenanthroline as a Fast and High-Performance Redox Mediator for Dye-
Sensitized Solar Cells, J. Phys. Chem. C 120 (2016) 9595−9603.
[25] Y. Saygili, M. Söderberg, N. Pellet, F. Giordano, Y. Cao, A. Belen Muñoz-García, S. M.
Zakeeruddin, N. Vlachopoulos, M. Pavone, G. Boschloo, L. Kavan, J.-E. Moser, M. Graetzel, A.
Hagfeldt, M. Freitag, Copper Bipyridyl Redox Mediators for Dye-Sensitized Solar Cells with High
Photovoltage, J. Am. Chem. Soc. 138 (2016) 15087−15096.
[26] J. Cong, D. Kinschel, Q. Daniel, M. Safdari, E. Gabrielsson, H. Chen, P. H. Svensson, L. Sun,
L. Kloo, Bis(1,1-bis(2-pyridyl)ethane)copper(I/II) as an efficient redox couple for liquid dye-
sensitized solar cells, J. Mater. Chem. A 4 (2016) 14550–14554.
25

ACCEPTED MANUSCRIPT
[27] J. Li, X. Yang, Z. Yu, G. G. Gurzadyan, M. Cheng, F. Zhang, J. Cong, W. Wang, H. Wang, X.
Li, L. Kloo, M.Wang, L. Sun, Efficient dye-sensitized solar cells with [copper(6,6′-dimethyl-2,2′-
bipyridine)2]2+/1+ redox shuttle, RSC Adv. 7 (2017) 4611–4615.
[28] M. Freitag, J. Teuscher, Y. Saygili, X. Zhang, F. Giordano, P. Liska, J. Hua, S.M.
Zakeeruddin, J.-E. Moser, M. Grätzel, A. Hagfeldt, Dye-sensitized solar cells for efficient power
generation under ambient lighting, Nat. Photonics 11 (2017) 372−378.
[29] A. Colombo, C. Dragonetti, M. Magni, D. Roberto, F. Demartin, S. Caramori, C.A. Bignozzi,
Efficient Copper Mediators Based on Bulky Asymmetric Phenanthrolines for DSSCs, ACS Appl.
Mater. Interfaces 6 (2014) 13945–13955.
[30] M. Magni, R. Giannuzzi, A. Colombo, M. P. Cipolla, C. Dragonetti, S. Caramori, S. Carli, R.
Grisorio, G. P. Suranna, C. A. Bignozzi, D. Roberto, M. Manca, Tetracoordinated Bis-
phenanthroline Copper-Complex Couple as Efficient Redox Mediators for Dye Solar Cells, Inorg.
Chem. 55 (2016) 5245−5253.
[31] A. Colombo, G. Di Carlo, C. Dragonetti, M. Magni, A. Orbelli Biroli, M. Pizzotti, D. Roberto,
F. Tessore, E. Benazzi, C. A. Bignozzi, L. Casarin, S. Caramori, Coupling of Zinc Porphyrin Dyes
and Copper Electrolytes: A Springboard for Novel Sustainable Dye-Sensitized Solar Cells, Inorg.
Chem. 56 (2017) 14189−14197.
[32] R. Grisorio, L. De Marco, C. Baldisserri, F. Martina, M. Serantoni, G. Gigli, G. P. Suranna,
Sustainability of Organic Dye-Sensitized Solar Cells: The Role of Chemical Synthesis, ACS
Sustainable Chem. Eng., 3 (2015) 770–777.
[33] R. Grisorio, L, De Marco, R. Giannuzzi, G. Gigli, G. P. Suranna, Molecular engineering of
largely π-extended metal-free sensitizers containing benzothiadiazole units: Approaching 10%
efficiency dye-sensitized solar cells using iodine-based electrolytes, Dyes and Pigments, 131 (2016)
282–292.
26

ACCEPTED MANUSCRIPT
[34] W. Sharmoukh, A. Attanzio, E. Busatto, T. Etienne, S. Carli, A. Monari, X. Assfeld, M. Beley,
S. Caramori, P.C. Gros, 2,5-Dithienylpyrrole (DTP) as a donor component in DTP–π–A organic
sensitizers: photophysical and photovoltaic properties, RSC Adv. 5 (2015) 4041−4050.
[35] J.-H. Yum, E. Baranoff, F. Kessler, T. Moehl, S. Ahmad, T. Bessho, A. Marchioro, E. Ghadiri,
J.-E. Moser, C.Yi, M.K. Nazeeruddin, M. Grätzel, A cobalt complex redox shuttle for dye-
sensitized solar cells with high open-circuit potentials, Nature Commun. 3 (2012) 361.
[36] A. K. Ichinaga, J. R. Kirchhoff, D. R. McMillin, C. O. Dietrich-Buchecker, P. A. Marnot, J.-P.
+
Sauvage, Charge-transfer absorption and emission of Cu(NN)2 systems, Inorg. Chem., 26 (1987)
4290–4292.
[37] N. A. Gothard, M. W. Mara, J. Huang, J. M. Szarko, B. Rolczynski, J. V. Lockard, L. X. Chen,
Strong Steric Hindrance Effect on Excited State Structural Dynamics of Cu(I) Diimine Complexes,
J. Phys. Chem. A, 116 (116) 1984−1992.
[38] N. Armaroli, Photoactive mono- and polynuclear Cu(I)–phenanthrolines. A viable alternative
to Ru(II)–polypyridines?, Chem. Soc. Rev., 30 (2001) 113−124.
[39] Y. Yao, M. W. Perkovic, D. P. Rillema, C. Woods, C., Structures and Properties of Copper(II)
and Copper(I) Complexes Containing an Ethane-Bridged Dimeric Phenanthroline Ligand, Inorg.
Chem., 31 (1992) 3956-3962.
[40] Md. A. Masood, D. J. Hodgson, Synthesis and Characterization of the Multidentate Ligand
2,9-Bis(N-pyrazolylmethyl)-l,10-phenanthroline (bpmp) and Its Copper(I) and Copper(II)
Complexes, Inorg. Chem., 32 (1993) 4839-4844.
[41] M. Magni, A. Colombo, C. Dragonetti, P. Mussini, Steric vs electronic effects and solvent
coordination in the electrochemistry of phenanthroline-based copper complexes, Electrochimica
Acta, 141 (2014) 324–330.
[42] R. A. Marcus, On the Theory of Oxidation‐Reduction Reactions Involving Electron Transfer. J.
Chem. Phys., 24 (1956) 966–978.
27

ACCEPTED MANUSCRIPT
[43] R. A. Marcus, On the Theory of Electron‐Transfer Reactions. VI. Unified Treatment for
Homogeneous and Electrode Reactions, J. Chem. Phys., 43 (1965) 679–701.
[44] L. Casarin, W.B. Swords, S. Caramori, C.A. Bignozzi, G. J. Meyer, Rapid Static Sensitizer
Regeneration Enabled by Ion Pairing, Inorg. Chem. 56 (2017) 7324-7327.
[45] J. G. Rowley, By. H. Farnum, S. Ardo, G. J. Meyer, Iodide Chemistry in Dye-Sensitized Solar
Cells: Making and Breaking I−I Bonds for Solar Energy Conversion, J. Phys. Chem. Lett., 1 (2010)
3132-3140.
[46] R. S. Gaddie, C. B. Moss, C. M. Elliott, Cyclic Voltammetric Study of Cobalt Poly-4-t-
butylpyridine Ligand Complexes on Glassy Carbon Electrodes: Electrolyte Dependence and
Mechanistic Considerations, Langmuir 29 (2013) 825-831.
[47] C. M. Elliott, S. Caramori, C. A. Bignozzi, Indium Tin Oxide Electrodes Modified with
Tris(2,2’-bipyridine-4,4’-dicarboxylic acid) Iron(II) and the Catalytic Oxidation of Tris(4,4’-di-tert-
butyl-2,2‘-bipyridine) Cobalt(II), Langmuir 21 (2005) 3022-3027.
[48] M. J. Scott, J. J. Nelson, S. Caramori, C. A. Bignozzi, C. M. Elliott, cis-Dichloro-bis(4,41-
dicarboxy-2,2-bipyridine)osmium(II)-Modified Optically Transparent Electrodes:ꢀApplication as
Cathodes in Stacked Dye-Sensitized Solar Cells, Inorg. Chem. 46 (2007) 10071-1078.
[49] S. Carli, L. Casarin, Z. Syrgiannis, R. Boaretto, E. Benazzi, S. Caramori, M. Prato, C. A.
Bignozzi, Single Walled Carbon Nanohorns as Catalytic Counter Electrodes for Co(III)/(II)
Electron Mediators in Dye Sensitized Cells, ACS Appl. Mater. Interfaces 8 (2016) 14604-14612.
28