, 629 (2011);
et al.Aswani Yella
Electrolyte Exceed 12 Percent Efficiency
Porphyrin-Sensitized Solar Cells with Cobalt (II/III)
This copy is for your personal, non-commercial use only.
clicking here.colleagues, clients, or customers by
, you can order high-quality copies for your
If you wish to distribute this article to others
The following resources related to this article are available online at
here.following the guidelines
can be obtained by
Permission to republish or repurpose articles or portions of articles
A correction has been published for this article at:
): October 10, 2012 www.sciencemag.org (this information is current as of
version of this article at:
including high-resolution figures, can be found in the online
Updated information and services,
can be found at:
Supporting Online Material
related to this article
A list of selected additional articles on the Science Web sites
3 articles hosted by HighWire Press; see:
This article has been
This article appears in the following
registered trademark of AAAS.
is a Science2011 by the American Association for the Advancement of Science; all rights reserved. The title
CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005.
(print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theScience
on October 10, 2012
Porphyrin-Sensitized Solar Cells with
Cobalt (II/III)–Based Redox Electrolyte
Exceed 12 Percent Efficiency
Aswani Yella,1Hsuan-Wei Lee,2Hoi Nok Tsao,1Chenyi Yi,1Aravind Kumar Chandiran,1
Md.Khaja Nazeeruddin,1Eric Wei-Guang Diau,3* Chen-Yu Yeh,2*
Shaik M Zakeeruddin,1* Michael Grätzel1*
The iodide/triiodide redox shuttle has limited the efficiencies accessible in dye-sensitized solar
cells. Here, we report mesoscopic solar cells that incorporate a Co(II/III)tris(bipyridyl)–based redox
electrolyte in conjunction with a custom synthesized donor-p-bridge-acceptor zinc porphyrin dye
as sensitizer (designated YD2-o-C8). The specific molecular design of YD2-o-C8 greatly retards
the rate of interfacial back electron transfer from the conduction band of the nanocrystalline
titanium dioxide film to the oxidized cobalt mediator, which enables attainment of strikingly
high photovoltages approaching 1 volt. Because the YD2-o-C8 porphyrin harvests sunlight
across the visible spectrum, large photocurrents are generated. Cosensitization of YD2-o-C8
with another organic dye further enhances the performance of the device, leading to a
measured power conversion efficiency of 12.3% under simulated air mass 1.5 global sunlight.
cost-effectiveness compared with silicon (Si)–
based photovoltaic devices (1). So far, the best
conversion efficiencies have been obtained with
ruthenium-based dyes used together with the
al excitation, the dye injects an electron into the
conduction band of a nanocrystalline film of a
wide band gap oxide, such as titanium dioxide
(TiO2), and is subsequently regenerated back
to the ground state by electrondonationfrom a
redox couple present in the electrolyte. In this
fashion, the DSC achieves the separation of light
harvesting and charge generation from charge
carrier transport, whereas all other known pho-
demands on the puritiy of the semiconductor—
99.9999% for solar grade Si—resulting in high
carrier transport achieved by the DSC has many
ye-sensitized solar cells (DSCs) have
recently received great attention be-
cause of their ease of fabrication and
do not have to undergo costly purification or
doping treatments. Although the DSC already
outperforms its competitors in ambient light and
indoor conditions, its validated solar-to-electric
power-conversion efficiency (PCE) under stan-
dard air mass 1.5(AM 1.5) reporting conditions
(1000 W/m2solar light intensity and 298 K) of
11.1% (2) is still a factor of 2 below that of Si
solar cells. (The air mass number expresses the
ratio of the path length of the solar light in the
atmosphere over that corresponding to vertical
position of the sun. At air mass 1.5, the sun is at
Because of an excessive loss of voltage dur-
ing the dye-regeneration reaction, the use of the
iodide/triiodide electrolyte as a redox shuttle lim-
to 0.8 Vand is thus a drawback of current DSC
embodiments. This has prevented substantial
gains in PCEoverthelast5 years(11).Tofurther
improve the PCE, development of redox media-
tors exhibiting higher reduction potentials than
that of I3–is warranted; Co(II/III)tris(bipyridyl)
complexes do have this property. However, such
one-electron, outer-sphere redox couples often
yield shorter electron lifetimes when used in the
DSC. This faster interfacial charge recombina-
tion,when compared with iodide/triiodide–based
redox electrolytes (12–18), lowers the photovolt-
age and reduces the efficiency of charge collec-
tion, decreasing the short circuit photocurrent
density (Jsc) and hence the PCE (19–21). Re-
cent evidence suggests that the introduction of
long-chain alkyloxy groups in the dye structure
may retard the unwanted charge recombination
process (22). Donor-p-bridge-acceptor (D-p-A)
sensitizers, endowed with such groups, recently
reached Vocvalues exceeding 0.8 V when used
ever, the PCEs of these devices remained in the
6.7 to 9.6% range because of their insufficient
solar light harvesting, resulting in low photocur-
rents (22–24). This rationale has inspired us to
synthesize a D-p-A zinc (Zn) porphyrin dye, des-
ignated YD2-o-C8, which absorbs light over the
whole visible range and is endowed with long-
chain alkoxy groups so as to impair interfacial
juntion with cobalt polypyridyl–based redox elec-
exceeding those obtained with today’s best ruthe-
Inspired by the important role that porphyrins
play in natural photosynthesis, researchers have
as sensitizers for DSCs, but conversion efficien-
cies obtained so far have largely remained below
8% (25–33). An exception is the class of D-p-A
porphyrin dyes, such as YD2, which has reached
a PCE of 11% when used with iodide/triiodide
redox electrolyte (34). A diarylamine group at-
tached to the porphyrin ring acts as an electron
donor, and an ethynylbenzoic acid moiety serves
as an acceptor, anchoring the dye to the titania
surface. The unique feature of these sensitizers
is that the porphyrin chromophore itself consti-
The judiciously tailored variant of YD2 that
we report here,YD2-o-C8 (Fig. 1), incorporates
two octyloxy groups in the ortho positions of
each meso-phenyl ring, producing a striking
1Laboratory for Photonics and Interfaces, Institute of Chemical
Sciences and Engineering, É cole Polytechnique Fédérale de
Lausanne, Lausanne-1015, Switzerland.2Department of Chem-
istry and Center of Nanoscience and Nanotechnology, Na-
tional Chung Hsing University, Taichung, Taiwan 402, ROC.
3Department of Applied Chemistry and Institute of Molecu-
lar Science, National Chiao Tung University, Hsinchu, Taiwan
*To whom correspondence should be addressed. E-mail:
firstname.lastname@example.org (M.G.); email@example.com
(S.M.Z.); firstname.lastname@example.org (E.W.-G.D); cyyeh@dragon.
Fig. 1. The molecular structures of the (left) YD2 and (right) YD2-o-C8 porphyrin dyes.
VOL 3344 NOVEMBER 2011
CORRECTED 2 DECEMBER 2011; SEE LAST PAGE
on October 10, 2012
amelioration of the photo-induced charge sep-
aration in DSCs using Co(II/III)tris(bipyridyl)–
11.9% have been achieved with this molecular
photovoltaic system, which produces a Vocof
965 mV, a Jscof 17.3 mA/cm2, and a fill factor
(FF) of 0.71 under standard AM 1.5 sunlight
at 995 W/m2intensity. The cosensitization of
D-p-A dye, coded Y123, yielded an efficiency
of 12.3% when used in conjunction with the
Co(II/III)tris(bipyridyl)–based redox electrolyte.
The PCE even exceeds 13% under AM 1.5 solar
light of 500 Wm−2intensity.
Molecular properties and photovoltaic per-
formance. The detailed synthetic procedure for
preparing the YD2-o-C8 porphyrin is described
in tetrahydrofuran (THF) solvent are shown in
fig. S1, and table S1 lists the corresponding
ima in the 400 to 500 nm and 550 to 750 nm
ranges, corresponding to the Soret and Q bands,
respectively, with similar molar absorption co-
efficients. Nevertheless, the Soret band of YD2-
o-C8 is slightly red-shifted, and its Q band is
narrower as compared with the YD2 spectrum.
This is ascribed to the electronic effect of the
two electron-donating dioctyloxy substituents
introduced in ortho-position of the meso-phenyl
ring. The emission behavior for the two sensitiz-
ers matches their absorption spectra, the fluores-
by 13 nm relative to that of YD2. The oxidation
potentials of the two dyes were determined by
means of cyclic voltammetry (CV). YD2-o-C8
shows two oxidation waves at half-wave poten-
tial (E1/2) = +0.82 and +1.37 V versus normal hy-
lower than the corresponding potentials of YD2
uents on the meso-phenyl rings.The 190-mVneg-
ative shift of the reduction potential of YD2-o-C8
corresponds to a lifting of the lowest unoccupied
tation of the same effect.
To gain further insight into the electron den-
calculations on the YD2-o-C8 porphyrin using
density functional theory (DFT) at the B3LYP/6-
31G(d) level [Spartan 08 package (Wavefunction,
Irvine, CA)] and compared them with the calcu-
lated electronic structure of the YD2 dye (fig. S3).
the highest occupied molecular orbital (HOMO)
and HOMO-1 is shared by the diphenylamine
donor moiety and the p-system of the porphy-
rin ring. Introduction of the strongly electron-
donating group (octyloxy) at the ortho-positions
of phenyl rings, attached at the meso-positions
of porphyrin core in YD2-o-C8, increases the
electronic density on the porphyrin p-system
over tert-butyl groups present at the HOMO and
HOMO-1 level of the YD2 dye. Thus, for YD2-
o-C8 there is a considerable electronic coupling
between the alkoxy groups and the porphyrin
core, lifting the LUMO level higher thanthatof
the YD2 dye. The predicted effect is an increase
of the HOMO-LUMO gap of YD2-o-C8 com-
with the experimental observations.
an electrolyte by using the Co(II/III)tris(bipyridyl)
consists of 0.165 M [CoII(bpy)3](B(CN)4)2,
tonitrile as a solvent. Details on the optimization
are presented in fig. S4. Using the Nernst equa-
tion and standard potential (Eo) = 0.57 V versus
NHE for the experimentally determined (23)
standard potential of the Co(II/III)tris(bipyridyl)
tetracyanoborate couple, the oxidation potential for
this electrolyte is derived as 0.535 V versus NHE.
Performance in a Co(II/III)tris(bipyridyl)–
mediated solar cell. Photocurrent density versus
voltage (J-V) is shown in Fig. 2A; the curves
were measured under standard photovoltaic re-
porting conditions (AM 1.5 global sunlight at
1000 W/m2and a temperature of 298 K) for cells
with YD2-o-C8 or YD2 sensitizers used in
conjunction with theAY1electrolyte.YD2-o-C8
gives a PCE of 11.9% compared with only 8.4%
for the reference dye YD2. The key observation
here is that a subtle modification of the porphyrin
structure induces strikingly large Jscand Vocim-
from 10% to full solar intensity (Table 1).
The incident monochromatic photon-to–
electric current conversion efficiency (IPCE) as
afunction of wavelength and J-V curves for the
same YD2 and YD2-o-C8 devices is presented
in Fig. 2. YD2-o-C8 shows impressively high
IPCE values over the whole visible wavelength
except for a narrow dip in the spectrum around
530 nm. The spectral response of the photocur-
rent obtained with the YD2 cell shows similar
features, with peaks at around 460 and 650 nm.
lower for YD2 as compared with YD2-o-C8,
despite its higher molar extinction coefficient,
which is in keeping with the lower short-circuit
photocurrent densities observed for YD2 in
Fig. 2A. The Jscvalues obtained from calculat-
ing the overlap integral of the IPCE spectrum
with the standard AM 1.5 global spectral solar
photon flux were 17.3 and 14.9 mA/cm2for YD2-
o-C8 and YD2, respectively. These figures agree
within 2% with the measured Jsc, showing that
true AM 1.5 solar emission is very small with
the solar simulator used in our experiments.
light-harvesting efficiency (LHE), the quantum
yield of electron injection from the excited sen-
sitizer into the TiO2conduction band (Finj), the
efficiency for dye regeneration (hreg), and the
IPCE = LHE Finjhreghcoll
The charge carrier collection efficiency is de-
(ttrans)and recombination (trec) of the conduction
Fig. 2. (A) Comparison
solar cells under full AM
current conversion effi-
ciencies as a function of
wavelength for the two
ness is a 5-mm transpar-
ent layer and on top a
5-mm scattering layer.
t i s
4 NOVEMBER 2011VOL 334
on October 10, 2012
band electrons injected into the nanocrystalline
hcoll= 1/(1 + ttrans/trec)(2)
We separately examined the four efficiency pa-
for the much higher IPCE values obtained with
The LHE of a 6-mm-thick transparent TiO2film
loaded with a monolayer of the sensitizer is
shown inFig.3.The LHEforthetwoporphyrins
is very similar, the Q-band for YD2-o-C8 being
slightly shifted to the blue with respect to that of
YD2, which matches the behavior of the ab-
near the Soret and Q bands, even for the 6-mm-
thick transparent TiO2film. The spectral do-
main where all the photons are captured by the
dye-sensitized mesoscopic TiO2film is further
broadened for the 10-mm-thick double layer
used in Fig. 2 because of the increase of the op-
tical path length by light scattering from the top
particle layer. This rules out that a change in the
LHE causes the difference in the IPCE and Jsc
values for the two porphyrins (34).
Dividing the IPCE by the LHE gives the
absorbed photon-to–electric current generation
efficiency (APCE), presenting the true quantum
yield for electric current generation from light.
The APCE values in the wavelength range from
420 to 700 nm are between 80 and 100%, and
60 to 90% for YD2-o-C8 and YD2, respectively
(Fig. 4). This indicates that the electron injec-
tion, dye regeneration, or charge carrier collec-
tion is less efficient for the latter than for the
former sensitizer when used in conjunction with
the Co(II/III)tris(bipyridyl) tetracyanoborate com-
plexes as redox couple.
We performed time-resolved luminescence
experiments to unravel any differences in the
dynamics of photo-induced electron injection
between the two porphyrins. In fig. S5, we com-
pare emission data in THF solution and for YD2-
o-C8– or YD2-covered nanocrystalline TiO2films
in contact with the same Co(II/III)tris(bipyridyl)–
based redox electrolyte used in the photovoltaic
experiments. In solution, the fluorescence life-
times for YD2-o-C8 and YD2 are 1.5 and 1.2 ns,
respectively, whereas the emission for both sen-
sitizers is strongly quenched in contact with
TiO2. Data fitting, including reconvolution of
the emission signal over the ~1 ns excitation pulse
length, indicates that the lifetime of the excited
singlet state of the two porphyrins in the adsorbed
state is at most 100 ps. These experiments do not
reveal any significant difference in the fluores-
cence kinetics of the two porphyrins. There is
also little doubt that the very rapid quenching
of luminescence leads in both cases to near-
quantitative charge injection from their excited
singlet state to the conduction band of the TiO2
nanocrystals, as implied by the very high APCE
values observed for YD2-o-C8 in Fig. 4 and the
near-unity IPCE values for YD2 obtained with
iodide/triiodide–based redox electrolytes.
The data obtained from intensity-modulated
photo-induced absorption (PIA) measurements
(41) shown in the top row of fig. S6 confirm the
occurrence of oxidative quenching after light
excitation of the two porphyrins adsorbed on the
nanocrystalline TiO2film. The absorption peaks
at 550, 800, and 1400 nm reveal the formation
of porphyrin radical cations by electron injection
from the excited state in the conduction band of
the titania. The two PIA spectra in the lower row
show that these features disappear completely in
the presence of the Co(II/III)tris(bipyridyl) electro-
lyte, indicating complete regeneration of the YD2
or YD2-o-C8 porphyrins through electron dona-
tion from the Co(II) form of the redox couple.
Because the LHE as well as the quantum
yield for electron injection and regeneration
are very high for both dyes, we reasoned that the
lower IPCE values for YD2 as compared with
YD2-o-C8 reflect differences in the efficiency
for carrier collection between the two porphyrin
Inhibition of back electron transfer. We used
transient photocurrent and photovoltage decay
measurements to compare the rates of interfacial
Fig. 3. LHE as a func-
tion of wavelength for
the YD2 and YD2-o-C8
porphyrins adsorbed at
Fig. 4. APCEasafunction
of wavelength for the YD2
and YD2-o-C8 porphyrin
adsorbed at the surface
of a 6-mm-thick nanocrys-
talline TiO2film in contact
550600 650 700
Table 1. Detailed photovoltaic parameters of the devices made with the dyes YD2 and YD2-o-C8 and
cobalt-based AY1 electrolyte at different light intensities. Pin, incident intensity of AM1.5 solar light.
VOL 3344 NOVEMBER 2011
on October 10, 2012
recombination of electrons from the TiO2con-
duction band to CoIIItris(bipyridyl) (42, 43). The
electron lifetime as a function of Vocfor YD2
and YD2-o-C8 is plotted in Fig. 5A. The Voc
was adjusted by varying the intensity of the bias
light impinging on the cell. The electron lifetime
is 2 to 10 times longer for YD2-o-C8– than for
YD2-sensitized TiO2films, the difference becom-
ing larger as the Vocincreases. One plausible ex-
electrons in the conduction band of TiO2(e–TiO2)
reacts faster with CoIIItris(bipyridyl) for films
loaded with the latter than with the former dye.
It appears that the specific molecular struc-
ture of YD2-o-C8—in particular, the presence of
the four octyloxy groups—reduces the recombi-
nation rate, most likely by inhibiting the access
by CoIIItris(bipyridyl) to the TiO2surface. The
distance dependence of the back reaction is in
accordance with the semiclassical electron trans-
fer theory (44, 45).
The lower rate of electron recapture by
CoIIItris(bipyridyl) forYD2-o-C8–sensitized nano-
crystalline TiO2films allows very high open-
circuit voltages to be realized, with this sensitizer
sacrificing short circuit photocurrent or fill factor.
maintained at lower light levels, down to 3 W/m2
solar intensity (fig. S7). The CoII/IIItris(bipyridyl)
electrolyte yields slopes of 90 mV/decade and
80 mV/decade, corresponding to ideality factors
o-C8 and YD2 dyes, respectively. This suggests
that surface-trapping states participate in the
back-electrontransferreaction,which is similar to
the findings by Hagfeldt and co-workers (22).
It may be argued that the higher Vocobserved
with YD2-o-C8 with respect to YD2 arises from
a larger dipole moment of the former as com-
pared with the latter sensitizer in the adsorbed
state, causing an upward shift of the conduction
band of TiO2. This would displace the trap-state
distribution function toward higher energies,
rendering the density of occupied states (DOS)
at a given forward bias voltage lower for YD2-
o-C8 than for YD2. To check this possibility,
the DOS of the films loaded with sensitizer was
determined from transient photocurrent decay
measurements (42, 43). The chemical capacitance
Cmfor the YD2-o-C8– and YD2-sensitized films
as a function of Vocare compared in Fig. 5B. Cm
is directly proportional to the DOS: Cm= q(e)
DOS, where q(e) is the charge of one electron.
The Cmvalues are very similar for the two sen-
sitizers, with small differences being noted only
at Voc> 0.8 V. This rules out any substantial
contribution of a conduction band shift to the
observed decrease in the back-electron trans-
with that of a I–/I3––based redox electrolyte using
again a (5+5) mm double-layer TiO2film. The
concentrations of I–and I3–in this electrolyte,
designated AY2, were identical to those of the
Co2+and Co3+tris(bipyridyl) in AY1; the other
additives, such as TBP and LiClO4, were also
maintained at the same levels, as was the ace-
and fig. S8 show that the photovoltaic perform-
ance of devices using the CoII/IIItris(bipyridyl)
electrolyte is far superior to that of the I–/I3––
based devices. The Vocand Jscvalues increase
by 193 mV and 2.3 mA/cm2, respectively, pro-
ducing a 58% gain of the PCE from 7.6 to 11.9%.
To evaluate the photovoltaic performance of the
newly designed YD2-o-C8 dye under conditions
that are optimal for I–/I3––based DSCs, a thicker
[1.0 M 1,3-dimethylimidazolium iodide (DMII),
0.03Miodine, 0.1Mguanidinium thiocyanate,
and 0.5 M tert-butylpyridine in a mixture of
valeronitrile/acetonitrile (15:85 v/v)]. Under these
voltaic parameters are in table S2). These results
Fig. 5. Electron (A) lifetime
and (B) capacitance deter-
mined with photocurrent and
photovoltage decay measure-
ments of devices with YD2
and YD2-o-C8 dyes.
0.64 0.68 0.72 0.76 0.80 0.84 0.88 0.92 0.96 1.00
0.65 0.700.75 0.80 0.850.900.95
Fig. 6. (A) J-V char-
acteristics of a YD2-o-
AM 1.5 global sunlight
sities. The molar YD2-o-
solution was 7. (B) Spec-
tral response of the IPCE
for YD2-o-C8 (red dots),
tized nanocrystalline TiO2
films (black squares)
Photocurrent density [mA/cm2]
4 NOVEMBER 2011VOL 334
on October 10, 2012
confirm that for YD2-o-C8, the CoII/IIItris(bipyridyl)
complex outperforms the I–/I3––based redox
electrolyte, even under optimal conditions for
Despite having excellent light-harvesting prop-
erties, the YD2-o-C8 dye lacks absorption in
the 480 to 630 nm range, as evidenced by the
IPCE spectra in Figs. 2B and 6B. The dip in the
green spectral region reduces the JSCand hence
the PCE. To avoid this loss, we used a cosensitizer
coded Y123 (a D-p-A dye whose structure is
shown in fig. S9) that possesses a complemen-
tary absorption spectrum to YD2-o-C8. Y123
exhibits an absorption maximum at 532 nm in
THF with an extinction coefficient of 53,000 M−1
cm−1(23). This coincides with the minimum in
the spectral IPCE response of the YD2-o-C8
dye, the absorption maximum of which is at
644 nm (e = 31,200 M−1cm−1).
Enhanced performance from adding a co-
sensitizer. The cosensitized films achieved bet-
ter photovoltaic performance than that of solar
cells using a single dye. The photocurrent J-V
curves measured at three different light inten-
sities, using the AY1 redox electrolyte, are shown
in Fig. 6A. The photovoltaic performance pa-
rameters are listed in Table 2. Increased short-
circuit current density and FF are observed for the
cosensitized films compared with YD2-o-C8 alone.
The cumulative increases of Jscand FF give rise
to an efficiency of 12.3% at AM 1.5 global full
sun using the CoII/IIItris(bipyridyl)–based redox
electrolyte. The PCE reached an even higher val-
ue of 13.1% at 509 W/m2solar intensity (50.9%
sun). These results were independently verified by
repeating the photovoltaic characterization of the
cells in the Photovoltaic Laboratory at the Institute
of Micro Technique (IMT), Neuchâtel, Switzerland.
The Wacom high-precision class AAA solar sim-
ulator system available at the IMT Photovoltaic
Laboratory very closely mimics the solar spec-
trum in the absorption range of the cosensitized
solar cells in the range of 350 to 750 nm. This
avoids any substantial spectral mismatch between
the simulated and true AM 1.5 solar light source.
Results shown in fig. S10 and table S3 fully
confirm the PCE measurements carried out in
our own laboratory.
The IPCE spectra of devices made with the
individual and combined dyes are shown in
Fig. 6B. The Jscvalue obtained from integrat-
ing the product of the IPCE spectrum with the
AM 1.5 global spectral solar photon flux was
18.3 mA/cm2. This value is greater than the mea-
sured Jscvalue (17.6 mA/cm2) at full sunlight.
The difference can be attributed to mass trans-
port limitations at full sunlight, which limit the
photocurrent. This effect is corroborated by the
photocurrent transient measurements by using
an on/off modulation of the incident light (46).
The data reported in fig. S11 show a spike in the
photocurrent, which reaches a lower stationary
value after a few seconds of illumination time.
Our goal is to eliminate this small Jscloss so as
to attain a strictly linear response of the photo-
current up to full solar intensity. Such linear
behavior already can be achieved by reducing
the porphyrin content in the YD2-o-C8/Y123–
containing staining solution from 7:1 to 4:1, as
shown in table S4. This indicates that the mass
transport limitation is also partly caused by
the bulky nature of the YD2-o-C8 dye. Because
the increase in Y123 content in the dye mixture
also reduces the Vocvalue, the linear photocur-
rent response does not contribute to the PCE at
The YD2-o-C8/Y123 cosensitized nano-
crystalline TiO2film exhibit an impressive pan-
chromatic photocurrent response over the whole
visible range, achieving incident photon-to-electron
conversion efficiencies of >90% in a large wave-
length domain below 700 nm. Although the co-
sensitization increases the Jsc, the Vocdecreases
by about 30 mV with respect to films sensitized
by YD2-o-C8 alone. Transient photo-voltage de-
cay measurements giving access to the electron
lifetime were used to examine the reason for this
decrease. The plots of electron lifetime versus
open-circuit potential in fig. S12 show that the
incorporation of Y123 reduced the lifetime of the
photo-generated electrons in the TiO2conduction
band compared with YD2-o-C8 alone.
Continued exposure of YD2-o-C8–sensitized
solar cells for 220 hours to full sunlight at 30°C
showed only a 10 to 15% decrease of the overall
efficiency over this extended light-soaking pe-
riod. This shows that the photosystem is robust
and that the small decline is probably caused by
losses of the volatile acetonitrile solvent.
The present results provide a fertile base for
further investigation, which will focus on achiev-
ing linear response of the photocurrent to full
sunlight, enabling PCE values above 13% to be
reached under standard reporting conditions. Fur-
ther challenges to be tackled are the use replace-
ment of the volatile electrolyte by nonvolatile
ionic liquid-based systems, as well as solid-state
hole conductors. Last, new anchoring groups will
be introduced in the porphyrin in order to avoid
any desorption on long-term heat and light ex-
posure, along with functional groups increasing
the near-infrared response of the sensitizer.
References and Notes
1. B. O’Regan, M. Grätzel, Nature 353, 737 (1991).
2. Y. Chiba et al., Jpn. J. Appl. Phys. 45, L638 (2006).
3. C.-Y. Chen et al., ACS Nano 3, 3103 (2009).
4. W. Zeng et al., Chem. Mater. 22, 1915 (2010).
5. Q. Yu et al., ACS Nano 4, 6032 (2010).
6. J. M. Kroon et al., Prog. Photovolt. Res. Appl. 15,
7. M. K. Nazeeruddin et al., J. Am. Chem. Soc. 127, 16835
8. Y. Chiba, A. Islam, R. Komiya, N. Koide, L. Y. Han,
Appl. Phys. Lett. 88, 223505 (2006).
9. Z. S. Wang, M. Yanagida, K. Sayama, H. Sugihara,
Chem. Mater. 18, 2912 (2006).
10. M. D. Wei et al., J. Mater. Chem. 16, 1287 (2006).
11. G. Boschloo, A. Hagfeldt, Acc. Chem. Res. 42, 1819 (2009).
12. M. Wang, N. Chamberland, L. Breau, J.-E. Moser,
R. H.-Baker, B. Marsan, S. M. Zakeeruddin, M. Grätzel,
Nat. Chem. 2, 385 (2010).
13. D. Li et al., Adv. Funct. Mater. 20, 3358 (2010).
14. T. Daeneke et al., Nat. Chem. 3, 211 (2011).
15. Y. Bai et al., Chem. Commun. (Camb.) 47, 4376 (2011).
16. B. A. Gregg, F. Pichot, S. Ferrere, C. L. Fields, J. Phys.
Chem. B 105, 1422 (2001).
17. S. Hattori, Y. Wada, S. Yanagida, S. Fukuzumi, J. Am.
Chem. Soc. 127, 9648 (2005).
18. Z. Zhang, P. Chen, T. N. Murakami, S. M. Zakeeruddin,
M. Grätzel, Adv. Funct. Mater. 18, 341 (2008).
19. H. Nusbaumer, S. M. Zakeeruddin, J. E. Moser,
M. Grätzel, Chemistry 9, 3756 (2003).
20. S. A. Sapp, C. M. Elliott, C. Contado, S. Caramori,
C. A. Bignozzi, J. Am. Chem. Soc. 124, 11215 (2002).
21. H. Nusbaumer, J. E. Moser, S. M. Zakeeruddin,
M. K. Nazeeruddin, M. Gratzel, J. Phys. Chem. B 105,
22. S. M. Feldt et al., J. Am. Chem. Soc. 132, 16714 (2010).
23. H. N. Tsao et al., ChemSusChem 4, 591 (2011).
24. Z. Difei et al., Environ. Sci. 4, 2030 (2011).
25. J. N. Clifford, G. Yahioglu, L. R. Milgrom, J. R. Durrant,
Chem. Commun. (Camb.) 12, 1260 (2002).
26. T. Hasobe et al., J. Am. Chem. Soc. 127, 1216 (2005).
27. T. Hasobe et al., J. Phys. Chem. B 109, 19 (2005).
28. M. Borgstrçm et al., J. Phys. Chem. B 109, 22928 (2005).
29. C.-W. Lee et al., Chem. Eur. J. 15, 1403 (2009).
30. A. Huijser, T. J. Savenije, A. Kotlewski, S. J. Picken,
L. D. A. Siebbeles, Adv. Mater. (Deerfield Beach Fla.) 18,
31. O. Hagemann, M. Jørgensen, F. C. Krebs, J. Org. Chem.
71, 5546 (2006).
32. G. M. Hasselman et al., J. Phys. Chem. B 110, 25430
33. S. Eu et al., J. Phys. Chem. C 111, 3528 (2007).
34. T. Bessho, S. M. Zakeeruddin, C. Y. Yeh, E. W. G. Diau,
M. Grätzel, Angew. Chem. Int. Ed. 49, 6646 (2010).
35. H.-P. Lu et al., J. Phys. Chem. C 113, 20990 (2009).
36. C.-P. Hsieh et al., J. Mater. Chem. 20, 1127 (2010).
37. S.-L. Wu et al., Environ. Sci. 3, 949 (2010).
38. J.-C. Chang, C.-J. Ma, G.-H. Lee, S.-M. Peng, C.-Y. Yeh,
Dalton Trans. 8, 1504 (2005).
39. H.-P. Lu et al., Phys. Chem. Chem. Phys. 11, 10270 (2009).
40. C.-L. Mai et al., Chem. Commun. (Camb.) 46, 809 (2010).
41. H. J. Snaith et al., Nanotechnology 19, 424003 (2008).
42. B. C. O’Regan, F. J. Lenzmann, J. Phys. Chem. B 108,
43. B. C. O'Regan, Bakker, J. Kroeze, H. Smit, P. Sommeling,
James R. Durrant, J. Phys. Chem. B 110, 17155 (2006).
44. P. Siders, R. A. Marcus, J. Am. Chem. Soc. 103, 741 (1981).
45. J. N. Clifford, E. Martínez-Ferrero, A. Viterisi,
E. Palomares, Chem. Soc. Rev. 40, 1635 (2011).
46. J. J. Nelson, T. J. Amick, C. M. Elliott, J. Phys. Chem. C
112, 18255 (2008).
Acknowledgments: E.W.-G.D. and C.Y.Y. acknowledge
the financial support from the National Science Council
of Taiwan and Ministry of Education of Taiwan. M.G.
thanks the European Research Council (ERC) for an
Advanced Research Grant (ARG 247404) funded under
the “Mesolight” project. Financial support under the
European community’s 7th FWP for project 227057
(INNOVASOL) and under a grant from the Dayton U.S. Air
force Research Laboratory is gratefully acknowledged.
Table 2. Detailed photovoltaic parameters of the devices made with AY1 electrolyte and YD2-o-C8 dye
cosensitized with Y123 dye at different light intensities.
VOL 3344 NOVEMBER 2011
on October 10, 2012
M.K.N. thanks World Class University program, Photovoltaic
Materials, Department of Material Chemistry, Korea
University, Chungnam 339-700, Korea, which is funded
by the Ministry of Education, Science and Technology
through the National Research Foundation of Korea
(R31-2008-000-10035-0). We thank R. Humphry-Baker
for fruitful discussions as well as optical measurements
and M. J. M. Bonnet-Eymard from the IMT Photovoltaic
Laboratory for assistance with the cell photovoltaic
Supporting Online Material
Materials and Methods
Figs. S1 to S12
Tables S1 to S4
References (47, 48)
13 June 2011; accepted 23 September 2011
Structural Dynamics of a Catalytic
Monolayer Probed by Ultrafast
2D IR Vibrational Echoes
Daniel E. Rosenfeld,* Zsolt Gengeliczki,* Brian J. Smith, T. D. P. Stack, M. D. Fayer†
Ultrafast two-dimensional infrared (2D IR) vibrational echo spectroscopy has proven broadly useful
for studying molecular dynamics in solutions. Here, we extend the technique to probing the interfacial
dynamics and structure of a silica surface-tethered transition metal carbonyl complex—tricarbonyl
(1,10-phenanthroline)rhenium chloride—of interest as a photoreduction catalyst. We interpret
the data using a theoretical framework devised to separate the roles of structural evolution and
excitation transfer in inducing spectral diffusion. The structural dynamics, as reported on by a
carbonyl stretch vibration of the surface-bound complex, have a characteristic time of ~150
picoseconds in the absence of solvent, decrease in duration by a factor of three upon addition
of chloroform, and decrease another order of magnitude for the bulk solution. Conversely,
solvent-complex interactions increase the lifetime of the probed vibration by 160% when solvent
is applied to the monolayer.
nologies, including industrial catalysis, chemical
sensors (2), fuel cells, and molecular electronics
(3). The functional groups terminating the mono-
layer determine the hydrophobicity, chemical re-
activity (4), and charge transfer properties (5) of
the interface, which are strongly influenced by lo-
cal structure and fast associated dynamics. Despite
a long-standing need, the tools to study structural
dynamics of interfacial molecules under chem-
ically relevant conditions have been lacking (6).
Commonly used microscopy and scattering tech-
niques provide information on the size, shape, and
electronic structure of particles and adsorbates,
but their time resolution is generally insufficient
to study molecular dynamics, and many only func-
tion under ultrahigh vacuum conditions (7).
The development of ultrafast infrared (IR)
spectroscopy over the past two decades has pro-
vided tools for the in-depth examination of the
dynamics and structure of bulk liquids, liquids
in nanoscopic environments, organic complexes,
biological macromolecules, and solids (8–11).
ailoring surface properties by deposit-
ing molecular monolayers (1) on various
solid substrates is critical to many tech-
IR techniques such as pump-probe absorption,
transient grating, and two-dimensional (2D) IR
vibrational echo spectroscopy have been used to
study spectral diffusion (9), vibrational relaxa-
tion (12), chemical exchange, (8, 10), and orien-
tational dynamics (11). The extension of these
techniques to surfaces and interfaces has been a
long-standing goal of the surface and ultrafast
spectroscopy communities (13).
Sum-frequency generation (SFG) and second-
harmonic generation (SHG) form the current
basis for vibrational spectroscopy of surfaces and
interfaces. Frequency-domain experiments pro-
vide important information on the molecular
orientation (14), vibrational coupling (15), and
hydrogen-bond network at interfaces (16), where-
as time-domain studies can probe reorientational
(17) and translational motions (18), thermal con-
ductance (19), vibrational relaxation (20), and
spectroscopic line broadening (21). The measure-
ment and quantitative interpretation of 2D IR
spectra of molecular adsorbates has previously
been limited to thick samples and attempts at 2D
IR-pump SFG-probe spectroscopy (upconverted
hole-burning) (22, 23). Upconverted two-pulse
vibrational echoes, which measure the homoge-
neous component (ultrafast motionally narrowed
dynamics) of the infrared absorption spectrum
but cannot study spectral diffusion (time depen-
dence of structural evolution), have been attempted
as well (21, 24). The frequency upconversion
in time-resolved SFG renders interpretation dif-
ficult because of the included Raman process
(23), which necessitates careful determination
of the time correlation functions measured in
the given beam geometry (25). Furthermore,
SFG-based techniques are inherently insensitive
because upconversion is inefficient. Techniques
relying on hole-burning methods have intrinsi-
cally lower time resolution and sensitivity and
produce convoluted spectra, whereas echo-based
methods suffer none of these drawbacks (26).
Here, we report on the application of an ultra-
fast 2D IR vibrational echo method to molecular
monolayers that overcomes all of these chal-
lenges, because there is no intrinsic tradeoff be-
tween time and frequency resolution, and the
associated heterodyne detection provides much
higher sensitivity (signal-to-noise ratio) than other
methods. This approach opens the way for the
quantitative understanding of the effect of im-
mobilization and solvent on the structural and
vibrational dynamics of molecular monolayers
on solid substrates.
We applied our technique to the study of a
silica-immobilized transition metal carbonyl com-
pound of interest as a photocatalyst. Immobilized
homogeneous catalysts are appealing because
they maintain high molecular specificity and
activity under mild conditions while preclud-
ing the need for expensive separation methods
(27–29). However, conditions such as the pres-
ence or absence of solvent can strongly affect
catalytic activity in presently unpredictable ways
(29, 30). Microscopic changes affecting molec-
ular reactivity should manifest themselves in the
dynamical characteristics of the system due to
the small energy differences among states. As a
first step toward resolving solvent-dependent vi-
brational dynamics that could ultimately assist
rational catalyst optimization, we have compared
spectral diffusion rates and vibrational lifetimes
of bare (monolayer/air interface) versus solvated
surface-bound complexes. In addition, the dy-
namics of the immobilized catalyst are compared
to the corresponding homogeneous catalyst in
class of complexes (L, heteroaromatic bidentate
ligand; X, halide/pseudohalide) that are under in-
for the reduction of CO2to CO or formate (30).
Other metal-ligand–based catalysts have been
immobilized (27), and the system depicted in
Fig. 1 represents a good model system for studying
Department of Chemistry, Stanford University, Stanford, CA
*These authors contributed equally to this work.
†To whom correspondence should be addressed. E-mail:
4 NOVEMBER 2011VOL 334
on October 10, 2012
1 Download full-text
CorreCtions & CLarifiCations
CorreCtions & CLarifiCations
www.sciencemag.org sCiEnCE erratum post date 2 deCemBer 2011
Research Article: “Porphyrin-sensitized solar cells with cobalt (II/III)–based redox electro-
lyte exceed 12 percent efficiency” by A. Yella et al. (4 November, p. 629). The Fig. 3 inset
legend is incorrect; the red line corresponds to YD2-o-C8, and the blue line corresponds to
YD2. Also, an error in the author affiliations was introduced in proofs. The correct affiliations
are “2Department of Chemistry and Center of Nanoscience and Nanotechnology, National
Chung Hsing University, Taichung, Taiwan 402, ROC. 3Department of Applied Chemistry and
Institute of Molecular Science, National Chiao Tung University, Hsinchu, Taiwan 300, ROC.”
The affiliations have been corrected in the HTML and PDF versions online.
Post date 2 December 2011
on October 10, 2012