Significant improvement of dye-sensitized solar cell performance by small structural modification in π-conjugated donor-acceptor dyes

Signiﬁcant Improvement of Dye-Sensitized Solar

Cell Performance by Small Structural Modiﬁcation

in π-Conjugated Donor–Acceptor Dyes

Stefan Haid, Magdalena Marszalek, Amaresh Mishra,* Mateusz Wielopolski,

Joël Teuscher, Jacques-E. Moser, Robin Humphry-Baker, Shaik M. Zakeeruddin,*

Michael Grätzel,* and Peter Bäuerle*

DSSCs based on ruthenium complexes

Two donor-π-acceptor (D-π-A) dyes are synthesized for application in dye-

sensitized solar cells (DSSC). These D-π-A sensitizers use triphenylamine

as donor, oligothiophene as both donor and π-bridge, and benzothiadiazole

(BTDA)/cyanoacrylic acid as acceptor that can be anchored to the TiO₂sur-

face. Tuning of the optical and electrochemical properties is observed by the

insertion of a phenyl ring between the BTDA and cyanoacrylic acid acceptor

units. Density functional theory (DFT) calculations of these sensitizers provide

further insight into the molecular geometry and the impact of the additional

phenyl group on the photophysical and photovoltaic performance. These dyes

are investigated as sensitizers in liquid-electrolyte-based dye-sensitized solar

cells. The insertion of an additional phenyl ring shows signiﬁcant inﬂuence on

the solar cells’ performance leading to an over 6.5 times higher efﬁciency (η =

8.21%) in DSSCs compared to the sensitizer without phenyl unit (η = 1.24%).

Photophysical investigations reveal that the insertion of the phenyl ring blocks

the back electron transfer of the charge separated state, thus slowing down

recombination processes by over 5 times, while maintaining efﬁcient electron

injection from the excited dye into the TiO₂-photoanode.

have broad absorption spectra extending

into the near-IR region and produce solar-

to-electrical energy conversion efﬁciencies

of up to 11.7% under AM1.5 irradiation.^[3–5]

The photoconversion efﬁciencies (PCE) of

metal-free organic dyes are somewhat lower

(≤10%) in comparison to the ruthenium-

based sensitizers, but they include advan-

tageous features, such as lower material

costs, better tunable absorptions, and band-

gaps by variation of the molecular structure,

which typically consists of a π-conjugated

donor-acceptor (D-A) system.^[6]In this

respect, various D-π-A-substituted organic

dyes have been prepared and used as sen-

sitizers in DSSCs.^[7–9]

One of the drawbacks of organic dyes

is their relatively narrow absorption spec-

trum, whose maxima usually remain in

the shorter wavelength region. The optimal

organic sensitizer for solar cell applications

should possess a broadly extended absorp-

tion spectrum ranging to the near-IR regime in order to attain

good overlap with the solar emission spectrum and to produce

large photocurrent responses. In addition, suitable energy levels

of the highest occupied molecular orbital (HOMO) and lowest

unoccupied molecular orbital (LUMO) of the dye are required to

match the iodide/triiodide redox potential and the conduction

band edge level of the TiO₂semiconductor electrode in order

to assure efﬁcient electron injection into the TiO₂conduction

band. For the latter process, the electron density distribution in

the dye’s HOMO and LUMO is also of crucial importance. One

way of extending the absorption range to enhance the photocur-

rent response is to reduce the HOMO–LUMO energy gap of the

sensitizer. For this purpose, various low bandgap sensitizers for

DSSCs have been prepared by implementing, e.g., the electron-

deﬁcient benzothiadiazole (BTDA) unit into the bridging frame-

work of D-π-A molecules.^[10–16]

1. Introduction

Dye-sensitized solar cells (DSSCs) are one of the most promising

environmentally friendly photovoltaic devices because of their low

material costs, ﬂexiblity, and easy manufacturing processes.^[1,2]

S. Haid, Dr. A. Mishra, Prof. P. Bäuerle

Institute of Organic Chemistry II and Advanced Materials

University of Ulm

Albert-Einstein-Allee 11, 89081 Ulm, Germany

E-mail: amaresh.mishra@uni-ulm.de; peter.baeuerle@uni-ulm.de

M. Marszalek, Dr. M. Wielopolski, Dr. J. Teuscher, Prof. J.-E. Moser,

Dr. R. Humphry-Baker, Dr. S. M. Zakeeruddin, Prof. M. Grätzel

Laboratory of Photonics and Interfaces

Institute of Chemical Sciences and Engineering

Ecole Polytechnique Fédérale de Lausanne (EPFL)

Station 6, CH-1015 Lausanne, Switzerland

In search of new metal-free sensitizers for DSSCs, we present

new D-A sensitizers 1 and 2 that are built up by a triarylamine-

bithiophene donor (D) and a BTDA-substituted cyanoacrylic

Email: shaik.zakeer@epﬂ.ch; michael.graetzel@epﬂ.ch

DOI: 10.1002/adfm.201102519

Adv. Funct. Mater. 2012, 22, 1291–1302

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1291

acid as acceptor and anchoring group (A). Dye 1 contains the

BTDA unit directly adjacent to the anchoring group, whereas

dye 2 contains “only” an additional phenyl ring between both

acceptor units. We investigate the inﬂuence of the structural

modiﬁcation of the sensitizer on optical, redox, and solar cell

performance. This subtle structural change in the sensitizer

induced a signiﬁcant inﬂuence on the DSSCs performance.

to bithiophene pinacol boronic ester 4 in a Suzuki–Miyaura

cross-coupling reaction to obtain 5 in an excellent yield of

96%. Lithiation of 5 with n-butyl lithium (n-BuLi) and subse-

quent quenching with tributyltin chloride (Bu₃SnCl) afforded

derivative 6 in quantitative yield. Hydrolysis of 4-bromo-7-

(bromomethyl)benzo[c][1,2,5]-thiadiazole 7 using potassium

carbonate in a dioxane/water mixture provided 4-bromo-7-

(hydroxymethyl)benzo[c][1,2,5]thiadiazole 8 in 64% yield. Fur-

ther oxidation of alcohol 8 with manganese dioxide (MnO₂)

in chloroform gave 7-bromobenzo[c][1,2,5]thiadiazole-4-

carbaldehyde 9 in 89% yield. Stille-type coupling of 6 and

bromo-derivative 9 or 4-(7-bromobenzo[c][1,2,5]thiadiazol-4-yl)

benzaldehyde 10^[17]gave intermediates 11 and 12 in yields of

86% and 92%, respectively. Final Knoevenagel condensation of

11 and 12 with cyanoacetic acid in the presence of ammonium

2. Results and Discussion

2.1. Synthesis

D-A dyes 1 and 2 were synthesized as shown in Scheme 1.

4-Bromo-N,N-bis[4-(hexyloxy)phenyl]aniline 3 was coupled

Scheme 1. Synthesis of dyes 1 and 2. a) Pd₂(dba)₃·CHCl₃, HP^tBu₃BF₄, 2 m aq. K₃PO₄(4 eq.), THF, room temperature (r.t.). b) (i) n -BuLi, THF, –78 °C

(ii) Bu₃SnCl. c) K₂CO₃, 1,4-dioxane/water, reﬂux, 1 h. d) MnO₂, CHCl₃, reﬂux, 3 h. e) Pd(PPh₃)₂Cl₂, THF, 70 °C, 15 h. f) Pd(PPh₃)₂Cl₂, THF, 75 °C,

4.5 h. g) cyanoacetic acid, NH₄OAc, CH₂Cl₂/CH₃CN, reﬂux.

1292 wileyonlinelibrary.com

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Funct. Mater. 2012, 22, 1291–1302

350 to 800 nm (Figure 1a). All data are summarized in Table 1.

The bands at 395 nm (dye 1) and 392 nm (dye 2) correspond to

the π–π^*transition of the conjugated system. The longest wave-

length absorption of dye 2 (λ_max= 515 nm) is assigned to a charge

transfer (CT) transition and is signiﬁcantly blue-shifted compared

to the CT-band of dye 1 (λ_max= 570 nm). Furthermore, the inten-

sity and molar extinction coefﬁcient for the CT transition of dye

2 (ε = 29,400 L mol⁻¹cm⁻¹) is increased by a factor of 1.6 com-

pared to dye 1 (ε = 18,900 L mol⁻¹cm⁻¹). The emission maxima

of dyes 1 and 2 can be found at 681 and 665 nm, respectively. Dye

2 showed a much larger Stokes shift (4379 cm⁻¹) in comparison

to 1 (2860 cm⁻¹), which is an indication of signiﬁcant structural

reorganization of 2 upon photoexcitation. The UV-vis spectra

of dyes 1 and 2 adsorbed on transparent TiO₂ﬁlms (Figure 1b)

showed a slight blue-shift of the CT band for dye 2 (λ_max

506 nm) compared to the solution spectra, while the shift is more

pronounced for dye 1 (λ_max= 540 nm). Hence, the steady-state

results clearly hint to a reduction of the overall π-conjugation in 2

due to a possible torsion of the additional phenyl ring.

2.3. Electrochemical Characterization

Cyclic voltammograms (CV) measured in dichloromethane

containing 0.1 m tetrabutylammonium hexaﬂuorophosphate

showed two reversible oxidation waves and one quasireversible

reduction wave for both sensitizers (Figure 2). The data are pre-

sented in Table 1. The ﬁrst oxidation potential (E⁰_ox1) for 1 and

2 was found at 0.20 and 0.11 V, respectively, (vs. Fc⁺/Fc) and are

assigned to the oxidation of the triphenylamine donor moiety.

The easier oxidation of 2 when compared to 1 is a further sup-

port for the weaker donor–acceptor interaction in dye 2 due

to probable torsion of the phenyl ring. The second oxidation

potential (E⁰_ox2) was observed at 0.60 V and 0.55 V for 1 and 2,

respectively, and is assigned to the oxidation of the conjugated

backbone. Due to the stronger acceptor character the reduction

potential for dye 1 is more positively shifted compared to dye 2

(ΔE_red= 0.19 V). The HOMO and LUMO energies determined

from the onset of the ﬁrst oxidation and reduction potential

were –5.22 and –3.62 eV vs. vacuum for dye 1 and –5.14 and

–3.47 eV for dye 2. Dyes 1 and 2 showed a bandgap (E_g^CV) of

1.60 and 1.67 eV, respectively, which are in good accordance to

the bandgaps determined from the optical studies (E_g^opt). The

calculated LUMO energy levels lie above the conduction band

edge of TiO₂(≈ –4.0 eV vs. vacuum), which should ensure suf-

ﬁcient driving force for electron injection from the dye into the

conduction band of the semiconductor.

Figure 1. a) UV-vis and normalized emission spectra of dyes 1 and 2

in dichloromethane ([c] = 10⁻⁵mol L⁻¹). Emission spectra were meas-

ured by excitation at 500 nm (1) and 460 nm (2) and were normalized

to their respective CT bands. b) UV-vis spectra of dyes 1 and 2 on TiO₂

ﬁlm (5 μm).

acetate as catalyst gave dye 1 in 86% yield and dye 2 in 38%

yield.

2.2. Spectroscopic Studies

The UV-vis absorption spectra of dyes 1 and 2 measured in

dichloromethane exhibited two absorption bands over a range of

Table 1. Optical and electrochemical data for dyes 1 and 2.

HOMO^a)

[eV]

LUMO^a)

[eV]

E_g

[eV]

E_g

[eV]

CVb)

optc)

Compound

E⁰

[V]

E⁰

[V]

E⁰

red

[V]

λ_max,abs

[nm]

ε_max

λ_max,em

[nm]

ox1

ox2

[L mol⁻¹cm⁻¹

]

395

570

392

515

26 400

18 900

38 000

29 400

681

665

0.20

0.11

0.60

0.55

1.60

1.67

−1.56

−1.75

−5.22

−3.62

1.67

1.98

−5.14

−3.47

^a)set Fc⁺/Fc E_HOMO= −5.1 eV; ^b)Calculated from the difference between the values of E_onset,

and E_onset,

;

^c)Estimated using the onset of the UV-vis spectra in

ox1

red1

dichloromethane.

Adv. Funct. Mater. 2012, 22, 1291–1302

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1293

sity distribution of the HOMO of both molecules is mainly

located at the donor part (oligothiophene and triphenylamine),

whereas the electron density of the LUMO is primarily located

at the BTDA and cyanoacrylic acid acceptor units and to a

small extent at the neighboring thiophene ring. Hence, both

orbitals provide sufﬁcient overlap between donor and acceptor

to guarantee a fast charge transfer transition. Thus, excitation

from the HOMO to the LUMO should lead to efﬁcient photoin-

duced electron transfer from the electron donating triarylamine

moiety to the TiO₂ﬁlm via the terminal cyanoacrylic acid. This

nicely follows the trends of the almost identical charge sepa-

ration rates for 1 and 2 as stated in the photophysical section

(see below). Signiﬁcant differences arise when considering the

cationic form of 1 and 2, as calculated using the unrestricted

Hartree–Fock method. The optimized cationic structure of 1

becomes completely planar with dihedral angles of ≈1° each

(Figure S1, Supporting Information). In contrast, the BTDA

unit and the adjacent phenyl rings in the radical cation of dye 2

are twisted out-of-plane by ≈48°. It should be noted that, in the

cationic form, the phenyl ring is twisted by additional 30° com-

pared to the neutral conformation (dihedral angle = 18°).

This deviation from planarity, as suggested by the large

Stokes-shift of 2 as compared to 1, has also an impact on the

orbital energies. In fact, the energies of the singly-occupied

molecular orbital (SOMO) are lowered to a higher extent in 2

than in 1. The SOMO of 2 (–6.80 eV) is 0.2 eV lower in energy

than the SOMO of 1 (–6.61 eV). This leads to the conclusion

that releasing an electron leads to a more stable radical cation

of 2 than that of 1. This ﬁnding corroborates well with the dif-

ference in solar cell performance of devices with 1 vs. 2 as dis-

cussed in the photovoltaic section. As shown below, the charge

recombination rates in 2 are remarkably slower due to a better

stabilization of its cationic form. Furthermore, the insertion of

the phenyl ring in 2 causes out-of-plane torsion with the adja-

cent units upon oxidation leading to an interruption of the

π-conjugation between donor and anchoring group. On the

contrary, the cationic form of 1 remains entirely planar and

the π-conjugation is preserved as shown in Figure S1. This fact

is clearly in favor of an intramolecular back-electron transfer

process as the deactivation pathway of the charge separated

state. In other words, the back-electron-transfer within the

donor-acceptor moieties of 1 is energetically preferred over the

charge recombination from TiO₂or a regen-

Figure 2. Cyclic voltammograms of 1 and 2 in DCM/TBAHFP (0.1 m),

[c] = 5 × 10⁻⁴mol L⁻¹, 295 K, scan rate = 100 mV s⁻¹, vs. Fc⁺/Fc.

2.4. Quantum Chemical Calculations

In order to gain insight into the structural properties, the elec-

tron density distribution of the frontier orbitals of both dyes

was analyzed by density functional theory (DFT) using the

B3LYP hybrid functional with 6–31G^*basis set (see Experi-

mental Section for details). First, the ground state geometries

of 1 and 2 were optimized using restricted Hartree–Fock condi-

tions. The calculations showed that the CT absorption bands of

these dyes are due to the HOMO → LUMO transition. The cal-

culated HOMO and LUMO energies were –4.96 and –3.36 eV,

respectively, for 1 and –4.81 and –3.01 eV for 2. In the neutral

form, both molecules do not exhibit signiﬁcant differences in

the electron density distribution of the orbitals (Figure 3).

In dye 1, the HOMO and LUMO orbitals are slightly more

delocalized, giving rise to a better conjugation between the

donor and acceptor moiety, which corroborates well with the

outcome of the UV-vis studies. On the other hand, in dye 2

the additional phenyl ring induces an out-of-plane torsion

of 18° with respect to the BTDA moiety. The electron den-

erating electrolyte.

2.5. Application in Dye-Sensitized Solar Cells

The current–voltage (J–V) characteristics of

solar cells sensitized with dyes 1 and 2 on

5 μm TiO₂ﬁlms with the volatile electrolyte

E1 measured under standard AM 1.5G con-

ditions (100 mW cm⁻²) are shown in Figure

S2 (Supporting Information). The electrolyte

compositions are presented in Table 2. Dye

2 sensitized cells provided an overall power

conversion efﬁciency (η) of 5.59% with a

high short circuit current density (J_SC) of

Figure 3. HOMO and LUMO electron density distributions for frontier orbitals of dyes 1 and

2 calculated using DFT methods (B3LYP/6-31G^*).

16.26 mA cm⁻², while dye 1 sensitized cells

1294 wileyonlinelibrary.com

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Funct. Mater. 2012, 22, 1291–1302

Table 2. Composition of electrolytes used for optimization of the per-

formance of the devices with dye 2. The components were dissolved

in a solvent mixture of acetonitrile and valeronitrile (85:15, v/v). DMII:

1,3-dimethylimidazolium iodide; GuNCS: guanidinium thiocyanate; TBP:

tert -butylpyridine; LiI: lithium iodide.

suppressed recombination processes while using the electro-

lyte with higher concentration of Li⁺. For the cells with lower

content of Li⁺, the electron lifetime is one order of magnitude

shorter than for the one with 100 mm Li⁺(see Figure S3, Sup-

porting Information).

The molecular structures of dye 1 and 2 differ only by an

additional phenyl ring placed between the BTDA unit and the

anchoring group in dye 2. Nevertheless, this small modiﬁca-

tion has a signiﬁcant impact on the photovoltaic performance.

In direct comparison with the above mentioned champion cell

of 2, using electrolyte E4, dye 1 exhibited a more than 5 times

lower J_SCand the resulting η was almost 6.6 times lower, as

shown in Figure 4a and Table 3. The difference in J_SCwas

further conﬁrmed by IPCE spectra. Both dyes showed a wide

spectral coverage of the 380–750 nm region (Figure 4b), but the

maximum IPCE at 600 nm for dye 1 is only 20% as compared

to >90% for dye 2.

Code

DMII [m]

I₂[m]

0.044

0.03

GuNCS [m]

TBP [m]

0.25

0.5

LiI [m]

0.1

0.03

0.5

0.05

0.1

0.03

0.5

showed a relatively poor performance with an overall η of

1.78% due to lower J_SCand open circuit potential (V_OC) values

(Table 3). The dramatically increased J_SCvalue for dye 2 was

also clearly seen in the incident photon-to-current conversion

efﬁciency (IPCE) spectra with ≈78% at 570 nm for the 2-based

device compared to only ≈28% at 550 nm for 1. The number

of the molecules of dye 1 and 2 desorbed from TiO₂ﬁlms of

the same surface area, thickness and porosity was found to be

1.20 × 10¹⁷and 1.07 × 10¹⁷, respectively. Dye loading experi-

ments showed that there is not much difference in the concen-

tration of the two dyes, indicating that the resulting absorptivity

is not the reason for the lower J_SCof the device with dye 1.

In order to optimize the conditions of the photovoltaic per-

formance for dye 2, we employed a thicker TiO₂layer with scat-

tering nanoparticles (8 μm transparent + 5 μm scattering) and

varied the LiI content in the electrolyte (coded as E1 to E4). This

allowed us to reach a very high J_SCvalue of 18.97 mA cm⁻²for

a LiI concentration of 1 m (E1). Unfortunately, V_OCand the ﬁll

factor decreased with increasing Li⁺concentration in the elec-

trolyte and on the TiO₂surface.^[18]It was found that the system

is very sensitive to the change of Li⁺concentration in the elec-

trolyte and the short-circuit current saturates even with small

amounts of Li⁺. At an optimal LiI concentration of 100 mm

(electrolyte E4), an excellent η of 8.21% was achieved. The

signiﬁcant difference in the photogenerated currents obtained

for electrolytes containing various concentrations of Li⁺may

be ascribed to an extended electron lifetime in the titania and

In order to explain this phenomenon, transient photovoltage

and photocurrent measurements were undertaken revealing

that the apparent electron lifetime in the case of the device

Table 3. Photovoltaic parameters of dye 1 and 2 adsorbed on nanoc-

rystalline TiO₂ﬁlms of different thicknesses. Electrolyte optimization

by varying the LiI content and its inﬂuence on the photovoltaic param-

eters of the cells sensitized with dye 2. Devices were made using single

(5 μm) and double layered TiO₂ﬁlm (8 μm transparent + 5 μm scat-

tering layer).

Dye Electrolyte LiI content

TiO₂ﬁlm

J_SC[mA

V_OC

[mV]

Fill

factor

[%]

[m]

thickness [μm] cm⁻²

]

6.04

423

593

489

640

688

655

558

0.70

0.57

0.74

0.69

0.73

0.65

0.55

1.78

5.59

1.24

8.21

7.19

7.26

5.91

16.26

3.40

0.1

8 + 5

18.47

14.26

17.52

18.97

Figure 4. a) J–V curves and b) IPCE spectra for devices sensitized with

dye 1 and 2 using a double-layered TiO₂ﬁlm (8 μm + 5 μm) and elec-

trolyte E4.

0.05

Adv. Funct. Mater. 2012, 22, 1291–1302

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1295

Figure 5. Apparent electron lifetimes of dye 1 and 2 measured with the

transient photovoltage technique.

Figure 7. Transient absorption data following the decay of the oxidized

state of dyes 1 and 2 adsorbed on a transparent TiO₂nanocrystalline ﬁlm

in presence and absence of redox couple.

with dye 2 exhibits typical exponential behavior when plotted

against charge density in the ﬁlm. It seems that for dye 1 an

asymptotic-like saturation limit exists for the amount of charges

injected into TiO₂. Above this limit, the recombination proc-

esses are vastly enhanced, thus reducing the lifetime of injected

electrons (Figure 5).

The time-correlated single photon counting (TCSPC) tech-

nique was used to study the excited state lifetimes of dyes 1 and

2 in solution and adsorbed on the TiO₂surface (Figure 6). Life-

times of the species in solution are in the nanosecond range:

5.6 ns for dye 1 and 7 ns for dye 2. In the case of molecules

adsorbed on the TiO₂layer, the lifetime decreased signiﬁcantly

as a result of the injection of the electron into the conduc-

tion band of the semiconductor. Dye 1 adsorbed on the TiO₂

substrate is characterized by an excited state lifetime, τ_TiO2of

296 ps, whereas dye 2 is slightly longer lived (τ_TiO2= 373 ps).

The estimated electron injection efﬁciencies for both dyes

were found to be fairly similar (dye 1: 94.8%; dye 2: 94.7%)

(Equation 1). The lack of a signiﬁcant difference between elec-

tron injection efﬁciencies, which cannot explain the difference

in device performance of 1 and 2, may turn the focus on the

charge collection efﬁciency or enhanced recombination proc-

esses as main reasons for the dramatically different perform-

ance of the two dyes in DSSCs.

τ_TiO

η_inj= 1 −

(1)

τ_solution

Nanosecond laser photolysis experiments focusing on the

decay of the oxidized state of the dye performed on TiO₂ﬁlms

covered with dye 1 and 2 in absence and presence of the redox

electrolyte have conﬁrmed 5 times faster recombination (half-

reaction time τ = 12 μs) in the case of dye 1 (Figure 7). Spectra

recorded in the presence of redox electrolyte exhibited a fast

decay (9 μs) due to the efﬁcient regeneration of the oxidized

dye molecules by iodide. Nevertheless, in the absence of the

redox couple, the dye radical cations can only be reduced by

the dye-injected electrons from TiO₂(recombination). In order

to ensure quantitative electron collection at the back contact

of the cell, this process is expected to be rather slow and non-

competing with the dye regeneration by I⁻. This requirement is

fulﬁlled in case of dye 2 and allows for the generation of high

current densities in the device. In contrast, dye 1 recombines

with electrons from TiO₂in the absence of a redox couple at

a time-scale comparable to that of the dye regeneration rate.

This undesirable behavior may indicate a facilitated electron

ﬂow from the TiO₂back to the dye radical cation. Furthermore,

this is in agreement with the planar molecular geometry of the

radical cation of dye 1 as compared to the twisted structure of

dye 2 radical cation as discussed above.

Figure 6. TCSPC ﬂuorescence histograms of dyes 1 and 2 dissolved in

acetonitrile and adsorbed on the surface of a TiO₂ﬁlm. Faster exponential

decays in case of dyes adsorbed on TiO₂are an indication of efﬁcient elec-

tron injection into the semiconductor. There is no signiﬁcant difference

between the injection yields of the two dyes.

1296 wileyonlinelibrary.com

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Funct. Mater. 2012, 22, 1291–1302

Figure 8. Differential absorption spectrum obtained upon femtosecond

pump-probe spectroscopy (λ_exc= 530 nm) of a ﬁlm of 1 on 5 μm thick

TiO₂with several time delays between 0 and 0.6 ps at room temperature

illustrating the radical cation formation with the ground-state transient

bleaching between 555 and 590 nm and the triarylamine cation (TA^•+)

beyond 680 nm.

Apart from the TCSPC and nanosecond ﬂash photolysis

studies, further insight into the excited state deactivation proc-

esses was provided by ultrafast transient absorption measure-

ments. Dyes 1 and 2 were studied on TiO₂ﬁlms in the absence

of a redox electrolyte. This permits to focus particularly on the

interactions between the dyes and TiO₂. 530 nm laser excita-

tion was used which directs the light energy mainly to the CT

absorption band of both dyes. At this wavelength, both dyes

exhibited similar extinction coefﬁcients, which allows for a good

comparison even at slight variations of ﬁlm thickness on TiO₂.

Immediately after laser excitation, the instantaneous formation

of the singlet excited states of the dyes was observed. Speciﬁ-

cally, the singlet excited states of 1 and 2 are characterized by

a maximum at 501 nm and transient bleach between 535 and

598 nm. After only 0.6 ps singlet-excited features of both dyes

transform into new bands (Figure 8).

This transduction of singlet-excited state energy is then iden-

tiﬁed by the shift of the minimum of the transient bleach from

555 to 584 nm and a broadening of the signal. The bleaching

is attributed to the ground-state bleaching due to the radical

cation formation of the triarylamine (TA^•+), which is repre-

sented by the appearance of a new absorption beyond 680 nm

(Figure 8). For both dyes, these features appear on a time-scale

of less than 2 ps. Hence, after a rapid transfer of singlet-excited

state energy, the formation of the charge separated state occurs

with equal rate constants for both dyes. Averaging ﬁrst-order

ﬁts of the time-absorption proﬁles leads to singlet-excited state

lifetimes of 0.95 ps for 1 and 0.79 ps for 2. This supports com-

parable charge-separation dynamics for both dyes and well

complies with the fact that the charge-injection efﬁciencies will

be equal when a redox electrolyte is present. The analysis of the

decay dynamics revealed the key difference between both dyes

(Figure S4, Supporting Information). For dye 1 the ground-state

bleaching and the radical cation signature are vanished after

300 ps, whereas in 2 they remain visible beyond this time scale.

Figure 9. First-order-ﬁts of the time-absorption proﬁles obtained from

the femtosecond pump-probe studies of 1 and 2 illustrating the dif-

ferent charge recombination dynamics of the radical cation signatures at

700 nm (a) and the ground-state bleaching at 580 nm (b).

The results clearly revealed the faster recombination process in

devices prepared using dye 1 compared to dye 2 devices.

In particular, in dye 1 the radical cation features decay with

a lifetime of 18 ps. On the contrary, in dye 2 they are persistent

on the time-scale of our experiment (1100 ps) and give rise

to a much longer lifetime of 526 ps (Figure 9). Conclusively,

charge recombination in 1 occurs almost 30 times faster than

in 2, which is in line with the fact that in the radical cationic

form of 1 the back-electron transfer is facilitated due to the pre-

served planarity of the π-conjugated system. In other words,

dye 1 recombines with electrons from TiO₂in the absence of

a redox couple due to a facilitated electron ﬂow from the TiO₂

back to the dye radical cation. On the contrary, the back elec-

tron transfer in 2 is hindered by the out-of-plane twist of the

additional phenyl ring upon oxidation. This leads to longer life-

times of the charge-separated state and thus may explain the

improved device performances.

Importantly, in this context the charge recombination life-

times do not match the lifetimes obtained by nanosecond laser

photolysis. The reason for this is that the laser intensities used

Adv. Funct. Mater. 2012, 22, 1291–1302

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1297

R_ct

η_coll

(2)

R_ct

R_t

The ﬁtting of the EIS measurements could not provide the

values required to calculate the charge collection efﬁciency for

the cell with dye 1 because a Gerischer-type impedance was

observed, where the transport resistance, R_t, and recombina-

tion resistance cannot be resolved explicitly.^[19]Nevertheless,

the presence of Gerischer impedance conﬁrms that the back

electron transfer reaction is faster than the transport of charge

carriers through the ﬁlm, which explains the very low J_SCmeas-

ured for the device with dye 1.

3. Conclusion

Two new thiophene-based D-π-A dyes incorporating a low

bandgap electron withdrawing benzothiadiazole (BTDA) unit

in the conjugated backbone were synthesized using Pd⁰-cata-

lyzed cross-coupling and Knoevenagel condensation reactions.

UV–vis spectroscopy, CV, and DFT calculations were performed

to examine the photovoltaic performance of these dyes. Incor-

poration of a BTDA unit close to the anchoring acceptor group

in dye 1 led to a bathochromic shift of the CT bands in the

UV-vis spectra in comparison to 2, in which the two acceptor

units are decoupled by a phenyl ring. However, the promising

capability of harvesting a broader range of photons by dye 1 did

not translate into improved cell performance. Cells with dye 1

exhibited ample injection efﬁciencies, but suffered from unusu-

ally fast recombination with the electrons from TiO₂, which was

reﬂected by the low PCEs. Insertion of a phenyl ring in order

to electronically separate the anchoring group and the BTDA

unit in 2 turned out to be a good solution to improve J_SC. The

additional phenyl ring in 2 seems to be twisted with respect

to the π-conjugated donor and BTDA part. Theoretical studies

revealed that upon formation of the radical cation of 2, the out-

of-plane torsion of the adjacent BTDA and cyanoacrylic acid

unit is enhanced leading to an interruption of the π-conjugation

between donor and anchoring group. The results prompt to the

conclusion that the formation of a stable dye 2 radical cation

and the interruption of the π-conjugation between the donor

and acceptor inhibit the back electron transfer. In particular,

the electron injection rate was barely affected, but the recom-

bination reaction was slowed down over 5 times. Thus, the efﬁ-

ciency of dye 2 (8.21%) was raised by a factor of 6.5 compared

to dye 1 (1.24%). It is of crucial importance to understand the

role of the dyes’ building blocks so as to improve the molecular

design in the future. Blocking this electron leak by structural

dye modiﬁcation with a twisted and electronically decoupling

unit may be a possible way to obtain high J_SCs and efﬁciencies

in DSSCs.

Figure 10. Charge transfer resistance as a function of the corrected

potential at the TiO₂/electrolyte interface measured by EIS in the dark

on the complete cells.

for the transient absorption studies exceed the laser power of

the nanosecond experiment by 2 orders of magnitude. As a

consequence, interfacial back electron ﬂow is facilitated due to

the high intensities and the generation of many more electrons

per laser pulse. This, on the other hand, leads to much faster

charge recombination dynamics as compared with the nano-

second experiments.

Electrochemical impedance spectroscopy (EIS) measurements

performed on complete cells in the dark revealed that the charge

transfer resistance of the TiO₂/electrolyte interface in the cell

sensitized with dye 1 is lower than that with dye 2 (Figure 10).

This may be the reason for faster recombination in the device

with dye 1, which is in accordance with our above-mentioned

results. Using Equation 2, where R_ctstands for a charge transfer

resistance and R_tis a transport resistance, the charge collection

efﬁciency (η_coll) for the best performing cell measured under

0.5 sun with dye 2 was found to be over 90% (Figure 11).

4. Experimental Section

Instruments and Measurements: NMR spectra were recorded on

a Bruker AMX 500 (¹H NMR: 500 MHz, ¹³C NMR: 125 MHz) or an

Avance400 spectrometer (¹H NMR: 400 MHz, ¹³C NMR: 100 MHz),

normally at 25 °C. Chemical shift values (δ) were expressed in parts

Figure 11. Charge collection efﬁciency for a cell with dye 2 as a function of

the corrected potential calculated from EIS measurement under 0.5 sun.

1298 wileyonlinelibrary.com

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Funct. Mater. 2012, 22, 1291–1302

per million using residual solvent protons (¹H NMR, δ_H= 7.26 for

CDCl₃and δ_H= 2.50 for DMSO-d₆; ¹³C NMR, δ_C= 77.0 for CDCl₃

and δ_C= 39.43 for DMSO-d₆) as the internal standard. The splitting

patterns were designated as follows: s (singlet), d (doublet), t (triplet),

q (quartet), and m (multiplet). The assignments were Ph-H (phenyl

protons), CHO (aldehyde protons), Th-H (thiophene protons), BTDA-H

(benzothiadiazole protons), HexOPh-H (4-hexyloxyphenyl protons).

Melting points were determined using a Büchi B-545 apparatus and

were not corrected. Elemental analyses were performed on an Elementar

VarioEL. Thin layer chromatography was carried out on aluminum

plates, precoated with silica gel, Merck Si60 F₂₅₄. Preparative column

chromatography was performed on glass columns packed with silica gel,

Merck Silica 60, particle size 40−63 μm. EI mass spectra were recorded

on a Varian Saturn 2000 GC-MS, CI mass spectra on a Finnigan MAT

SSQ-7000 and matrix-assisted laser desorption ionization time-of-ﬂight

(MALDI-TOF) mass spectra on a Bruker Daltonics Reﬂex III.

Materials: Tetrahydrofuran (THF) (Merck) was dried under reﬂux over

sodium/benzophenone (Merck). dichloromethane (CH₂Cl₂), chloroform

(CHCl₃), n-hexane, ethyl acetate, acetonitrile (MeCN), dioxane, and

methanol (MeOH) were purchased from Merck and distilled prior to

use. All synthetic steps were carried out under an argon atmosphere

(except Knoevenagel condensations). 2-Isopropoxy-4,4,5,5-tetramethyl-

1,3,2-dioxaborolane was purchased from Aldrich and was dried over

molecular sieve (4 Å) prior to use. Ammonium acetate (NH₄OAc) and

potassium carbonate were purchased from Merck, Pd₂(dba)₃·CHCl₃,

Pd(PPh₃)₂Cl₂, HP^tBu₃BF₄, n-BuLi (1.6 mol L⁻¹in hexane) were purchased

from Acros, and cyanoacetic acid was purchased from ABCR. The

potassium phosphate (K₃PO₄) solution was prepared by dissolving

potassium phosphate (ABCR) in deionized water and was degassed

prior to use. 4-Bromo-N,N-bis[4-(hexyloxy)phenyl]aniline 3,^[20]4-bromo-

7-(bromomethyl)benzo[c][1,2,5]-thiadiazole 7,^[21]2,2′-bithiophene,^[22]

and 4-(7-bromobenzo[c][1,2,5]thiadiazol-4-yl)benzaldehyde 10^[23]were

prepared according to literature procedures.

6.6 Hz, 4H, -OCH₂-), 1.72–1.65 (m, 4H, -OCH₂CH₂-), 1.44–1.36 (m, 4H,

-OCH₂CH₂CH₂-), 1.32–1.27 (m, 8H, -CH₂CH₂CH₃), 0.87 (t, J = 7.0 Hz,

6H, -CH₃).¹³C NMR (100 MHz, DMSO-d₆): δ = 155.41, 148.10, 142.50,

139.52, 136.56, 134.22, 128.30, 126.84, 126.04, 125.05, 124.98, 124.79,

123.60, 122.84, 119.12, 115.48, 67.63, 30.98, 28.69, 25.19, 22.05, 13.88.

MS (MALDI-TOF): m/z: [M⁺] 609.6 (calcd. for C₃₈H₄₃NO₂S₂609.3).

Elemental analysis: calc. (%) for C₃₈H₄₃NO₂S₂: C, 74.84; H, 7.11; N,

2.30; found: C, 74.70; H, 7.09; N, 2.27.

5-(N,N-Bis(4-hexyloxyphenyl)-4-aminophenyl)-5′-(tri-n-butylstannyl)-2,2′-

bithiophene (6): Bithiophene 5 (601 mg, 0.99 mmol) was dissolved in

dry THF (10 mL) and the solution was cooled to –78 °C. n-BuLi (1.6 m

in hexane, 0.63 mL, 1.01 mmol) was added dropwise and the solution

turned dark green. It was stirred for 1 h at this temperature, then

tributyltin chloride (360 mg, 1.06 mmol) was added. The reaction mixture

was allowed to slowly warm to r.t. and was stirred at r.t. overnight. Water

(30 mL) was added and the mixture was extracted with dichloromethane

(DCM) (2 × 30 mL). The combined organic layers were washed with

water, dried over Na₂SO₄, and the solvent was removed. The crude

stannyl compound 6 (943 mg, quant.) was obtained as a brownish oil. It

was directly used in the next step without further puriﬁcation. H NMR

(400 MHz, CDCl₃): δ = 7.39 (d, ³J = 8.8 Hz, 2H, Ph-H), 7.28 (d, ³J = 3.2 Hz,

1H, Th-H), 7.10 (d, ³J = 4.0 Hz, 1H, Th-H), 7.08–7.04 (m, 6H, HexOPh-H

+ ThH), 6.91 (d, J = 8.8 Hz, 2H, Ph-H), 6.85–6.81 (m, 4H, HexOPh-H),

3.94 (t, ³J = 6.4 Hz, 4H, -OCH₂-), 1.82–1.75 (m, 4H, -OCH₂CH₂-),

1.63–1.55 (m, 6H, -CH₂-), 1.51–1.45 (m, 4H, CH₂-), 1.43–1.31 (m, 14H,

-CH₂-), 1.15–1.10 (m, 6H, -CH₂-), 0.92 (t, ³J = 7.2 Hz, 6H, -CH₃), 0.91 (t,

³J = 7.2 Hz, 9H, -CH₃). ¹³C NMR (100 MHz, CDCl₃): δ = 155.62, 148.30,

143.10, 143.08, 140.48, 136.30, 136.11, 135.60, 126.71, 126.60, 126.24,

126.18, 124.47, 124.30, 122.22, 120.46, 115.33, 68.31, 31.63, 29.36,

28.97, 27.27, 25.79, 22.63, 14.05, 13.67, 10.92. MS (MALDI-TOF): m/z:

[M⁺] 899.5 (calcd. for C₅₀H₆₉NO₂S₂Sn 899.4).

4-Bromo-7-(hydroxymethyl)benzo[c][1,2,5]thiadiazole (8): 4-Bromo-7-

(bromomethyl)-benzo[c][1,2,5]thiadiazole 7 (2.67 g, 8.67 mmol) and

potassium carbonate (3.60 g, 26.0 mmol) were suspended in a dioxane/

water mixture (50 mL, 1:1 v/v) and the mixture was reﬂuxed for 1 h. The

solvents were removed in vacuo and the residue was partitioned between

2-(2,2′-Bithien-5-yl)-4,4,5,5-tetramethyl-1,3,2-dioxaborolane (4): 2,2′-

Bithiophene (2.00 g, 12.0 mmol) was dissolved under argon in dry THF

(30 mL). At –78 °C, n-butyllithium (1.6 m in hexane, 7.52 mL, 12.0 mmol)

was added dropwise within 20 min. The greenish solution was allowed

to warm to r.t. within 75 min. 2-Isopropoxy-4,4,5,5-tetramethyl-1,3,2-

dioxaborolane (2.50 mL, 12.3 mmol) was slowly added at –78 °C,

the solution was allowed to warm to r.t. within 2 h and was stirred at

r.t. for 6 h. The reaction mixture was poured into a saturated NH₄Cl

solution (30 mL) and was extracted with diethyl ether (3 × 30 mL). The

combined organic phases were washed with water, dried over Na₂SO₄,

and the solvent was evaporated to provide boronic ester 4 as viscous oil

(3.31 g, 94%). The product was dried in high vacuum and directly used

in the next step without further puriﬁcation. ¹H NMR (400 MHz, CDCl₃):

m hydrochloric acid (60 mL) and dichloromethane (60 mL). The

aqueous phase was extracted with DCM (2 × 50 mL) and the combined

organic phases were dried over Na₂SO₄. Removal of the solvent afforded

2.13 g of crude product. It was puriﬁed by column chromatography

(silica; hexane/ethyl acetate 4:1). Product 8 was obtained as an off-white

powder (1.36 g, 5.57 mmol, 64%). Mp. 139 °C. ¹H NMR (400 MHz,

CDCl₃): δ = 7.82 (d, ³J = 7.2 Hz, 1H, BTDA-H-5), 7.47 (dt, ³J = 7.2 Hz, ⁴J

= 1.0 Hz, 1H, BTDA-H-6), 5.13 (d, J = 1.0 Hz, 2H, CH₂OH), 2.52 (s br,

1H, CH₂OH). ¹³C NMR (100 MHz, CDCl₃): δ = 153.47, 153.08, 133.27,

132.01, 126.94, 113.21, 61.82. MS (CI): m/z (%): 247 ([M⁺]+2, 94), 245

([M⁺]+H, 100), 229 (87), 227(83), 217(18), 215(18). Elemental analysis:

calc. (%) for C₇H₅BrN₂OS: C, 34.30; H, 2.06; N, 11.43; found: C, 34.55;

H, 2.22; N, 11.30.

δ = 7.55 (d, J = 3.6 Hz, 1H, Th-H), 7.28–7.24 (m, 3H, Th-H), 7.05–7.02

(m, 1H, Th-H), 1.38 (s, 12H, -CH₃). ¹³C NMR (100 MHz, CDCl₃): δ =

144.01, 137.88, 137.26, 127.86, 124.93, 124.91, 124.33, 124.31, 84.14,

24.74. MS (EI): m/z (%): 292 (M⁺, 100).

7-Bromobenzo[c][1,2,5]thiadiazole-4-carbaldehyde (9): 4-Bromo-7-

(hydroxy-methyl)benzo[c][1,2,5]thiadiazole 8 (1.00 g, 4.08 mmol) and

manganese(IV) oxide (1.43 g, 16.4 mmol) were suspended in chloroform

(40 mL) and the mixture was ﬁrst stirred at r.t. for 16 h then reﬂuxed

for 3 h. The mixture was ﬁltered, the ﬁltrate was evaporated, and the

residue was puriﬁed by column chromatography (silica; hexane/ethyl

acetate 4:1). The pure aldehyde 9 (884 mg, 89%) was obtained as a

slightly yellowish solid. Mp. 192 °C. ¹H NMR (400 MHz, CDCl₃): δ =

5-(N,N-Bis(4-hexyloxyphenyl)-4-aminophenyl)-2,2′-bithiophene (5): Trip-

henylamine 3 (834 mg, 1.59 mmol) and boronic ester 4 (550 mg,

1.88 mmol) were dissolved under argon in dry THF (12 mL) and the

solution was degassed. Pd₂(dba)₃(29.8 mg, 32.5 μmol, 2 mol%) and

HP^tBu₃BF₄(18.6 mg, 62.2 μmol, 4 mol%) were added, and the solution

was degassed again. An aqueous solution of K₃PO₄(2 m, 3.2 mL,

6.4 mmol) was added and the mixture turned from dark red to green.

It was stirred at r.t. overnight. The reaction mixture was poured into

a saturated NH₄Cl solution (15 mL) and extracted with DCM (3 ×

15 mL). The combined organic phases were dried over Na₂SO₄and the

solvent was removed in vacuo. The resulting brown oil was puriﬁed by

column chromatography (silica; hexane/DCM 9:1 to 6:4). Bithiophene 5

(930 mg, 96%) was obtained as a greenish yellow solid. Melting point

(Mp.) 61–64 °C.¹H NMR (400 MHz, DMSO-d₆): δ = 7.47 (dd, ³J = 5.2

Hz, ⁴J = 0.8 Hz, 1H, Th-H), 7.44 (m, 2H, Ph-H), 7.27 (dd, ³J = 3.6 Hz, ⁴J

= 1.2 Hz, 1H, Th-H), 7.26 (d, ³J = 3.6 Hz, 1H, Th-H), 7.23 (d, ³J = 3.6 Hz,

10.74 (s, 1H, CHO), 8.09 (d, J = 7.6 Hz, 1H, BTDA-H-5), 8.04 (d, J =

7.6 Hz, 1H, BTDA-H-6).¹³C NMR (100 MHz, CDCl₃): δ = 188.34, 153.99,

152.28, 132.16, 131.68, 126.79, 121.89. MS (CI): m/z (%): 245 ([M⁺]+2,

100), 243 ([M⁺]+H, 99). Elemental analysis: calc. (%) for C₇H₃BrN₂OS:

C, 34.59; H, 1.24; N, 11.52; found: C, 34.40; H, 1.26; N, 11.39.

7-(5-(N,N-Bis(4-hexyloxyphenyl)-4-aminophenyl)-2,2′-bithien-5′-yl)

benzo[c]-[1,2,5]thiadiazole-4-carbaldehyde (11): In a Schlenk tube, stannyl

compound

6 (805 mg, 0.716 mmol) and 7-bromobenzo[c][1,2,5]

thiadiazole-4-carbaldehyde 9 (148 mg, 0.609 mmol) were dissolved under

argon in dry THF (15 mL) and the solution was degassed. Pd(PPh₃)₂Cl₂

(13.0 mg, 18.5 μmol, 3 mol%) was added and the solution was degassed

1H, Th-H), 7.07 (dd, J = 5.2 Hz, J = 3.6 Hz, 1H, Th-H), 7.01 (m, 4H,

HexOPh-H), 6.89 (m, 4H, HexOPh-H), 6.75 (m, 2H, Ph-H), 3.92 (t, J =

Adv. Funct. Mater. 2012, 22, 1291–1302

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1299

again. It was stirred at 70 °C for 15 h. The reaction mixture was poured

into water (30 mL) and extracted with DCM (3 × 30 mL). The combined

organic layers were dried over Na₂SO₄and the solvent was removed.

The crude product was puriﬁed by column chromatography (silica;

DCM:hexane 3:7 to 100:0, product eluted with DCM:EtOAc 9:1). Aldehyde

(E)-3-(4-(7-(5′-(4-(bis(4-(hexyloxy)phenyl)amino)phenyl)-2,2′-bithien-

5-yl)benzo[c][1,2,5]thiadiazol-4-yl)phenyl)-2-cyanoacrylic acid (2): Aldehyde

12 (150 mg, 0.18 mmol), cyanoacetic acid (302 mg, 3.55 mmol) and

NH₄OAc (13.2 mg, 0.17 mmol) were dissolved in a DCM/MeCN

mixture (20 mL, 2:1 v/v) and the deeply red solution was reﬂuxed for

45 h.The DCM was evaporated and a dark solid precipitated. It was

ﬁltered, washed with ethanol, and dried. It was further puriﬁed by

column chromatography (silica; aldehyde eluted with DCM, product

with DCM/MeOH 9:1). The product still contained some cyanoacetic

acid, therefore it was washed thoroughly with ethanol and dried in high

vacuum to obtain 62 mg (0.07 mmol, 38%) of dye 2 as deeply purple

solid. ¹H NMR (400 MHz, DMSO-d₆): δ = 8.36 (s, 1H, C = CH), 8.24 (d,

³J = 8.2 Hz, 2H, Ph-H), 8.18–8.13 (m, 4H, Ph-H + Th-H + BTDA-H), 8.03

(d, ³J = 7.6 Hz, 1H, BTDA-H), 7.46 (d, ³J = 8.4 Hz, 2H, HexOPh-H), 7.42

(d, ³J = 3.7 Hz, 1H, Th-H), 7.38 (d, ³J = 3.5 Hz, 1H, Th-H), 7.31 (d, ³J

= 3.5 Hz, 1H, Th-H), 7.02 (d, ³J = 8.6 Hz, 4H, HexOPh-H), 6.90 (d, ³J

= 8.6 Hz, 4H, HexOPh-H), 6.75 (d, ³J = 8.4 Hz, 2H, HexOPh-H), 3.93

11 (406 mg, 86%) was obtained as a purple solid. H NMR (400 MHz,

CDCl₃): δ = 10.68 (s, 1H, CHO), 8.23–8.18 (m, 2H, BTDA-H + Th-H),

7.93 (dd, ³J = 7.6 Hz, ⁵J = 1.6 Hz, 1H, BTDA-H), 7.39 (d, ³J = 8.4 Hz, 2H,

Ph-H), 7.26–7.25 (m, 2H, Th-H), 7.11 (d, ³J = 3.6 Hz, 1H, Th-H), 7.07 (d,

³J = 8.8 Hz, 4H, HexOPh-H), 6.91 (d, J = 8.4 Hz, 2H, Ph-H), 6.86–6.82

(m, 4H, HexOPh-H), 3.94 (t, ³J = 6.6 Hz, 4H, -OCH₂-), 1.82–1.74 (m,

4H, -OCH₂CH₂-), 1.51–1.43 (m, 4H, -OCH₂CH₂CH₂-), 1.37–1.32 (m,

8H, -CH₂CH₂CH₃), 0.92 (t, J = 7.2 Hz, 6H, -CH₃).¹³C NMR (100 MHz,

CDCl₃): δ = 188.42, 155.73, 153.72, 152.19, 148.71, 145.12, 142.26,

140.14, 136.35, 134.22, 132.79, 132.76, 131.20, 126.83, 126.27, 125.75,

125.32, 125.13, 124.26, 123.23, 122.42, 119.96, 115.29, 68.23, 31.59,

29.30, 25.75, 22.61, 14.05. MS (MALDI-TOF): m/z: [M⁺] 771.7 (calcd. for

C₄₅H₄₅N₃O₃S₃771.3). Elemental analysis: calc. (%) for C₄₅H₄₅N₃O₃S₃: C,

70.01; H, 5.87; N, 5.44; found: C, 69.87; H, 5.86; N, 5.34.

(t, J = 6.3 Hz, 4H, -OCH₂-), 1.73–1.66 (m, 4H, -OCH₂CH₂-), 1.45–1.37

(m, 4H, -OCH₂CH₂CH₂-), 1.35–1.26 (m, 8H, -CH₂CH₂CH₃), 0.88 (t, J =

(E)-2-Cyano-3-(7-(5-(N,N-bis(4-hexyloxyphenyl)-4-aminophenyl)-2,2′-

bithien-5′-yl)benzo[c][1,2,5]thiadiazol-4-yl)acrylic acid (1): Aldehyde 11 (165

mg, 0.214 mmol), cyanoacetic acid (363 mg, 4.27 mmol) and NH₄OAc

(14.7 mg, 0.191 mmol) were dissolved in a DCM/MeCN mixture (30 mL,

2:1 v/v) and the deeply purple solution was reﬂuxed for 7 h. The almost

black reaction mixture was poured into water (30 mL) and extracted

with DCM. The combined organic phases were dried over Na₂SO₄.

After removal of DCM a dark solid precipitated, which was collected

and washed thoroughly with ethanol. The product contained some

aldehyde, therefore it was further puriﬁed by column chromatography

(silica; aldehyde eluted with DCM, product with DCM:MeOH 9:1 to 1:1).

6.5 Hz, 6H, -CH₃). HRMS (MALDI-TOF) m/z: [M⁺] 914.29877 (calcd. for

C₅₄H₅₀N₄O₄S₃914.29887).

Optical and Cyclic Voltammetry Measurements: Optical measurements

were carried out in 1 cm cuvettes with Merck Uvasol grade solvents

and absorption spectra were recorded on a Perkin Elmer Lambda

19 spectrometer. Transparent TiO₂ﬁlms (5 μm) screen printed onto

a ﬂuorine-doped tin oxide (FTO) substrate were immersed in the

0.3 mm dye solutions in chlorobenzene for 4 h. The UV-vis spectra

were measured with a CARY 5 UV-Vis-NIR spectrophotometer (Varian,

Inc.). Cyclic voltammetry experiments were performed with a computer-

controlled Autolab PGSTAT30 potentiostat in a three-electrode single-

compartment cell with a platinum working electrode, a platinum wire

counter electrode and Ag/AgCl reference electrode. All potentials were

internally referenced to the ferrocene/ferrocenium couple.

Computational Details: All molecular orbital calculations were carried

out using the Gaussian 09 suite^[24]of programs with the B3PW91^[25]and

B3LYP^[26]hybrid functionals and the 6-31G^*[27]basis set. First, the ground

state geometries of 1 and 2 were optimized using the restricted Hartree–

Fock method. To ensure that the resulting structures represented the

global minimum on the potential energy surface the geometries were

computed using the B3LYP functional and the semiempirical AM1^*

level^[28]as implemented in the VAMP 10.0 software package.^[29]The

resulting deviation of the dihedral angles between the BTDA and phenyl

ring was less than 1°, which assured the global minimum. The long

hexyloxy chains, which have no signiﬁcant impact on the frontier orbitals

of the chromophores, were replaced with methoxy-substituents in order

to accelerate the convergence of optimizations.

Device Fabrication: Screen-printed double layers of TiO₂particles were

used as photoelectrodes in this study. For transparent ﬁlms a 5 μm thick

layer of 20 nm-sized TiO₂particles was printed on the ﬂuorine doped

SnO₂(FTO) conducting glass electrode. Double layered ﬁlms consisted

of a 8 μm thick transparent layer of 20 nm-sized TiO₂covered with a

5 μm thick scattering layer of 400 nm sized TiO₂particles. The porosity

was evaluated as 67% for the 20 nm TiO₂transparent layer and 42% for

the scattering layer, determined from BET measurements. After sintering

at 500 °C and cooling to 80 °C, the sintered TiO₂electrodes were

sensitized by dipping for 5 h in the respective dye solutions (0.3 mm of

the dye with 2 mm of chenodeoxycholic acid, CDCA, in chlorobenzene),

and then assembled using a thermally platinized FTO/glass counter

electrode. The working and counter electrodes were separated by a

25 μm thick hot melt ring (Surlyn, DuPont) and sealed by heating. The

cell’s internal space was ﬁlled with electrolyte using a vacuum pump. The

composition of the electrolytes is given in the Table 2. The electrolyte-

injecting hole on the thermally platinized FTO glass counter electrode

was ﬁnally sealed with a Surlyn sheet and a thin glass cover by heating.

Cells were equipped with self-adhesive antireﬂective foil that acted as

well as UV cut-off ﬁlter.

Dye 1 (155 mg, 86%) was obtained as a deeply purple solid. H NMR

(400 MHz, DMSO-d₆): δ = 8.84 (s, 1H, C=CH), 8.57 (d, ³J = 7.6 Hz, 1H,

BTDA-H), 8.20 (d, J = 7.6 Hz, 1H, BTDA-H), 8.12 (d, J = 3.2 Hz, 1H,

Th-H), 7.39 (d, ³J = 8.0 Hz, 2H, Ph-H), 7.35–7.32 (m, 2H, Th-H), 7.24

(d, ³J = 2.8 Hz, 1H, Th-H), 7.00 (d, ³J = 8.4 Hz, 4H, HexOPh-H), 6.88 (d,

³J = 8.4 Hz, 4H, HexOPh-H), 6.71 (d, J = 8.0 Hz, 2H, Ph-H), 3.91 (t, J

= 6.0 Hz, 4H, -OCH₂-), 1.72–1.65 (m, 4H, -OCH₂CH₂-), 1.44–1.35 (m,

4H, -OCH₂CH₂CH₂-), 1.33–1.24 (m, 8H, -CH₂CH₂CH₃), 0.87 (t, ³J =

7.2 Hz, 6H, -CH₃). HRMS (MALDI-TOF) m/z: [M⁺] 838.26747 (calcd. For

C₄₈H₄₆N₄O₄S₃838.26757).

4-(7-(5′-(4-(Bis(4-(hexyloxy)phenyl)amino)phenyl)-2,2′-bithien-5-yl)

benzo[c][1,2,5]thiadiazol-4-yl)benzaldehyde (12): In

a Schlenk tube,

4-(7-bromobenzo[c][1,2,5]thiadiazol-4-yl)benzaldehyde 10 (94.0 mg,

0.29 mmol) and stannylated bithiophene 6 were dissolved under argon in

dry THF (10 mL) and the solution was degassed. Pd(PPh₃)₂Cl₂(6.60 mg,

9.40 μmol, 3 mol%) was added and the solution was degassed again.

Afterwards, it was stirred at 75 °C for 4.5 h. The deeply red reaction

mixture was poured into water (20 mL) and extracted three times with

DCM. The combined organic layers were washed with water (20 mL),

dried over Na₂SO₄and the solvent was removed. The crude product

was puriﬁed by column chromatography (ﬂash silica; DCM:hexane

gradient elution 1:1 to 100:0, product eluted with pure DCM). Aldehyde

12 (229 mg, 0.27 mmol, 92%) was obtained as a red solid. ¹H NMR

(400 MHz, CDCl₃): δ = 10.10 (s, 1H, CHO), 8.15 (d, ³J = 8.4 Hz, 2H,

Ph-H), 8.09 (d, J = 4.0 Hz, 1H, Th-H), 8.03 (d, J = 8.4 Hz, 2H, Ph-H),

7.91 (d, J = 7.5 Hz, 1H, BTDA-H), 7.79 (d, J = 7.5 Hz, 1H, BTDA-H),

7.40 (d, ³J = 8.7 Hz, 2H, HexOPh-H), 7.25 (d, ³J = 3.8 Hz, 1H, Th-H), 7.23

(d, J = 3.8 Hz, 1H, Th-H), 7.12–7.06 (m, 5H, Th-H + HexOPh-H), 6.91

(d, J = 7.9 Hz, 2H, HexOPh-H), 6.84 (d, J = 8.8 Hz, 4H, HexOPh-H),

3.94 (t, J = 6.6 Hz, 4H, -OCH₂-), 1.82–1.75 (m, 4H, -OCH₂CH₂-), 1.51–

1.43 (m, 4H, -OCH₂CH₂CH₂-), 1.38–1.33 (m, 8H, -CH₂CH₂CH₃), 0.92

(t, ³J = 7.0 Hz, 6H, -CH₃). ¹³C NMR (100 MHz, CDCl₃): δ = 191.73, 155.72,

153.65, 152.53, 148.59, 144.36, 143.09, 140.27, 139.77, 137.15, 135.72,

134.75, 130.72, 129.90, 129.66, 128.99, 128.85, 127.33, 126.78, 126.24,

125.64, 125.12, 124.74, 124.01, 122.35, 120.13, 115.33, 68.28, 31.60,

29.32, 25.75, 22.60, 14.02. MS (MALDI-TOF) m/z: [M⁺] 847.7 (calcd. for

C₅₁H₄₉N₃O₃S₃847.3). Elemental analysis: calc. (%) for C₅₁H₄₉N₃O₃S₃: C,

72.22; H, 5.82; N, 4.95; found C, 72.36; H, 5.84; N, 4.96.

Photovoltaic Characterization: A 450 W xenon light source (Oriel, USA)

was used to characterize the solar cells. The spectral output of the lamp

1300 wileyonlinelibrary.com

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Funct. Mater. 2012, 22, 1291–1302

was matched in the region of 350–750 nm with the aid of a Schott K113

Tempax sunlight ﬁlter (Präzisions Glas & Optik GmbH, Germany) so as

to reduce the mismatch between the simulated and true solar spectra

to less than 2%. The current–voltage characteristics of the cell under

these conditions were obtained by applying external potential bias to the

cell and measuring the generated photocurrent with a Keithley model

2400 digital source meter (Keithley, USA). A similar data acquisition

system was used to control the incident photon-to-current conversion

efﬁciency (IPCE) measurement. Under computer control, light from a

300 W xenon lamp (ILC Technology, USA) was focused through a Gemini-

180 double monochromator (Jobin Yvon Ltd., UK) onto the photovoltaic

cell under test. The devices were masked to attain an illuminated active

area of 0.159 cm².

Dye Loading: Transparent TiO₂ﬁlms (4 μm) screen printed on a non-

conducting microscope slide were immersed in the solutions identical to

the ones used for cell preparation (0.3 mm dye powder + 2 mm of CDCA

in chlorobenzene) and kept in the dark for 4 h. Then the ﬁlms were rinsed

with chlorobenzene and placed in a solution of tetrabutylammonium

hydroxide (TBAOH) in DMF (0.5 g per 50 mL) for overnight desorption.

The UV-vis spectra were measured using TBAOH/DMF solution as a

blank on a CARY 5 UV-Vis-NIR spectrophotometer (Varian, Inc.).

Photocurrent and Photovoltage Transient Measurements: Transient

decays were measured under white light bias with superimposed red light

perturbation pulses (both light sources were light-emitting diodes (LEDs)).

The voltage dynamics were recorded by using a Keithley 2400 source

meter. Varying the intensity of white light bias allowed the estimate of

recombination rate constant (and thus apparent electron lifetime) at

different open-circuit potentials by controlling the concentration of the

free charges in TiO₂. Red perturbation pulses were adjusted to a very low

level in order to maintain single-exponential voltage decay.

double-stage noncollinear optical parametric ampliﬁer (NOPA) and to

produce a white light continuum in a sapphire plate or 387 nm UV light

by second harmonic generation of the laser output in a thin β-barium

borate (BBO) crystal. The NOPA was pumped by 200 μJ pulses at a

central wavelength of 775 nm and the excitation wavelength was tuned

to 530 nm to generate pulses of approximately 10 μJ. The output pulses

of the NOPAs were compressed in a SF10-glass prism pair compressor

down to a duration of less than 50 fs (full width at half maximum

(fwhm)). Iris diaphragms were used to decrease the pulse energy down

to a few microjoules for the pump and to less than 1 μJ for the probe

beam. Transient spectra were measured using a white light continuum

(WLC) for probing. The latter was generated from pulses (energy <10 μJ)

focused into a 2 mm thick sapphire plate. The monoﬁlament white beam

was collimated using a 90° off-axis paraboloid mirror and steered to

the sample using only reﬂective optics. A smooth monotonic spectral

distribution between 480 and 720 nm was obtained by controlling the

pump energy with an iris and a variable density ﬁlter. The polarization

between the pump and probe beams was controlled using a coherent

zero-order half waveplate and experiments were carried out in the magic

angle (54.7°) conﬁguration. Pump and probe beams were directed

parallel to each other toward a 60° off-axis paraboloid mirror that

focused them into the sample. A broadband membrane beam splitter

was placed before the sample to split the probe beam into reference and

signal arms. Both were collected by a lens system and directed into two

Andor Shamrock 163 imaging spectrographs. The absorbance change

was calculated using the ratio between data obtained with and without

the pump pulses reaching the sample and corrected for ﬂuctuations in

the WLC intensity using the reference beam spectra.

Photoinduced Absorbance: The PIA spectra of the various cells were

recorded over a wavelength range of approximately 500 to 1500 nm

following an (on/off) photomodulation using a 9 Hz square wave

emanating from a blue LED. White probe light from a halogen lamp was

used as illumination source.

Time-Correlated Single Photon Counting: Steady-state excited state

emission was measured with a FluoroLog-322 (Horiba) spectrometer

equipped with a 450 W Xe arc lamp. Time-correlated single photon

counting experiment was conducted using the above mentioned setup

with additional FluoroHub (Horiba) unit with TBX-04 photomultiplier as

a detector. For measurement a NanoLED pulsed laser-diode emitting at Supporting Information

406 nm was used. In case of measurement of the dye adsorbed on the

TiO₂surface (transparent, 3 μm) the setup was switched to the front-

face detection mode.

Supporting Information is available from the Wiley Online Library or

from the author.

Electrochemical Impedance Spectroscopy: The electrochemical

measurements were conducted on the complete devices in the dark at

20 °C. A sinusoidal potential perturbation of an amplitude of 10 mV Acknowledgements

was applied over a frequency range 1 MHz to 0.1 Hz (Autolab PG30)

S.H. and M.M. contributed equally to this work. A.M., S.H., and P.B. thank

for a potential bias varying between 100 and 660 mV with a 20 mV step.

Obtained spectra were resolved using the transmission model line.

Laser Studies: The nanosecond laser ﬂash photolysis technique was

applied to dye-sensitized, 8 μm-thick, transparent TiO₂mesoporous

ﬁlms deposited on normal ﬂint glass. Pulsed excitation (λ = 505 nm, 7 ns

pulse duration, 30 Hz repetition rate) was carried out by a Powerlite 7030

frequency-doubled Q-switched Nd:YAG laser (Continuum, Santa Clara,

California, USA). The laser beam output was expanded by a planoconcave

lens to irradiate a large cross-section of the sample, whose surface was

kept at a 30° angle to the excitation beam. The laser ﬂuence on the

sample was kept at a low level (30 μJ cm⁻²per pulse) to ensure that,

on average, less than one electron was injected per nanocrystalline TiO₂

particle on pulsed irradiation. The probe light, produced by a continuous

wave xenon arc lamp, was ﬁrst passed through a monochromator tuned

at 650 nm, various optical elements, the sample, and then through a

second monochromator, before being detected by a fast photomultiplier

tube (Hamamatsu, R9110). Data waves were recorded on a DSA 602A

digital signal analyser (Tektronix, Beaverton, Oregon, USA). Satisfactory

signal-to-noise ratios were typically obtained by averaging over 1500

laser shots.

the German Ministry of Education and Research (BMBF) in the frame of

project OPEG 2010. M.G. thanks the Swiss National Science Foundation

for the ﬁnancial support under the Indo-Swiss Joint Research Programme

(ISJRP) grant. M.W. and J.E.M. gratefully acknowledge support from the

NCCR MUST program of the Swiss NSF. The authors would like to thank

Pascal Comte for providing TiO₂layers used for the study.

Received: October 19, 2011

Published online: January 27, 2012

[1] M. Grätzel, Nature 2001, 414, 338.

[2] B. O’Regan, M. Grätzel, Nature 1991, 353, 737.

[3] M. Grätzel, Acc. Chem. Res. 2009, 42, 1788.

[4] Q. Yu, Y. Wang, Z. Yi, N. Zu, J. Zhang, M. Zhang, P. Wang, ACS

Nano 2010, 4, 6032.

[5] C.-Y. Chen, M. Wang, J.-Y. Li, N. Pootrakulchote, L. Alibabaei,

C.-H. Ngoc-le, J.-D. Decoppet, J.-H. Tsai, C. Grätzel, C.-G. Wu,

S. M. Zakeeruddin, M. Grätzel, ACS Nano 2009, 3, 3103.

[6] H. Zollinger, Color Chemistry: Syntheses, Properties, and Applications

of Organic Dyes and Pigments, 3rd ed., VHCA and Wiley-VCH, Zürich

and Weinheim, 2003.

Femtosecond Transient Absorption: Time-resolved transient absorption

measurements by the pump-probe technique used a compact CPA-2001,

1 kHz, Ti:sapphire-ampliﬁed femtosecond laser (Clark-MXR), with a pulse

width of about 120 fs and a pulse energy of 1 mJ at a central wavelength

of 775 nm. The output beam was split into two parts for pumping a

[7] A. Mishra, M. K. R. Fischer, P. Bäuerle, Angew. Chem. Int. Ed. 2009,

48, 2474.

Adv. Funct. Mater. 2012, 22, 1291–1302

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1301