A trisheteroleptic cyclometalated RuII sensitizer that enables high power output in a dye-sensitized solar cell

Article abstract of DOI:10.1002/anie.201104275

A trisheteroleptic cyclometalated ruthenium complex (see picture) bearing aliphatic substituents was synthesized, characterized, and evaluated in a dye-sensitized solar cell (DSSC). This thiocyanate-free dye, which has favorable structural elements to exh

Full text of DOI:10.1002/anie.201104275

Communications
DOI: 10.1002/anie.201104275
Solar Cells
A Trisheteroleptic Cyclometalated RuII Sensitizer that Enables High
Power Output in a Dye-Sensitized Solar Cell**
Paolo G. Bomben, Terry J. Gordon, Eduardo Schott, and Curtis P. Berlinguette*
The low embodied energy and high power-conversion effi-
ciency (h) over disparate light intensities renders the dye-
sensitized solar cell (DSSC)[1,2] a promising alternative to
conventional photovoltaic technologies.[3] Significant pene-
tration of the DSSC into the photovoltaic market, however, is
hindered predominantly by the long-term stability of dyes and
electrolytes under practical conditions.[4–6] The instability of
champion (i.e., h > 10%) dyes (which, until recently,[7] all
were derivatives of [Ru(dcbpy)2(NCS)2] (N3; dcbpy = 4,4’-
dicarboxy-2,2’-bipyridine)[2]) in the DSSC is caused primarily
by desorption of the dyes from the surface and/or liberation of
the NCSÀ ligands from the metal centre.[5,6] While the rate of
dye desorption from TiO2 can be manipulated by replacing
Scheme 1. Molecular design elements that can be manipulated to render
high-performance Ru-based DSSC dyes with a) NCSÀ groups and b) cyclo-
metalating ligands. The HOMO is localized at the metal and the anionic
À
the CO2H moiety with other anchoring groups, this strategy
ligands in both cases as indicated, but is more easily attenuated in the
typically compromises electron injection into the TiO2.[8] An
case of the cyclometalated complexes.[10]
alternative approach is to replace the dcbpy ligands that
comprise N3 with bidentate ligands bearing aliphatic sub-
stituents (e.g., Scheme 1a), which serve to hinder water from
reaching the surface to hydrolytically cleave the TiO2–dye
ester linkage.[9] These groups provide the additional benefit of
suppressing recombination between the electrolyte and the
electrons in TiO2, thus leading to higher efficiencies (Sche-
me 1a).[2]
injection into the TiO2.[10,11,15–18] This scenario leaves open the
opportunity to replace one dcbpy with a bidentate ligand
capable of suppressing recombination and enhancing the
optical properties as per the aforementioned protocol
(Scheme 1).[2,19] While we recently demonstrated synthetic
access to trisheteroleptic Ru sensitizers (e.g., 1 and 2;
Scheme 2),[20] we learned that removing the acid linkers
À
Chemical strategies for avoiding the labile Ru NCS bond
have been realized recently;[10,11] indeed, we[12] and others[13,14]
have now documented remarkably high h values for DSSCs
containing NCSÀ-free RuII sensitizers. Cyclometalated Ru
complexes such as [RuII(dcbpy)2(ppy)]1+ (ppy = 2-phenylpyr-
idine) provide a versatile platform in this respect because:
1) the highest occupied molecular orbital (HOMO) is
extended over the metal and anionic ring thus enabling its
modulation through judicious installation of substituents at
[15]
À
the R2 site in Scheme 1b; and 2) the low-lying excited
states, which contain orbital character that resides on the p*
framework of the dcbpy ligand(s), are poised for electron
Scheme 2. Trisheteroleptic cyclometalated Ru dyes 1-3.[20]
[*] P. G. Bomben, T. J. Gordon, Prof. C. P. Berlinguette
Department of Chemistry, and the Centre for Advanced Solar
Materials, University of Calgary
raises the HOMO level of the sensitizer to potentially
compromise dye regeneration. (The HOMO level of the
sensitizer must lie lower in energy than the IÀ/I3À redox couple
that resides at approximately + 0.5 V vs. normal hydrogen
electrode (NHE).[21] Although the HOMO of 1 lies at
+ 0.70 V vs. NHE and therefore meets this criterion,[20]
champion Ru-based sensitizers all have oxidation potentials
higher than ca. + 0.9 V.[2]) We therefore set out to overcome
this potential shortcoming by introducing strongly electron-
2500 University Drive NW, Calgary AB, T2N 1N4 (Canada)
E-mail: cberling@ucalgary.ca
E. Schott
Departamento de Ciencias Quꢀmicas
Universidad Andres Bello, Repfflblica 275, Santiago (Chile)
[**] This work was financially supported by the Canada Natural Science
and Engineering Research Council (NSERC), Canada Research
Chairs, Canadian Foundation for Innovation and Alberta Innovates.
Detailed experimental information including computational meth-
ods can be found in the Supporting information for this article,
which is available on the WWW under http://dx.doi.org/10.1002/
anie.201104275.
À
withdrawing CF3 substituents to the cyclometalating ligand
to accommodate efficient dye regeneration. These design
elements led to the preparation of 3—a RuII complex devoid
10682
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 10682 –10685

of NCSÀ groups and bearing aliphatic substituents that can
generate high efficiencies in a conventional DSSC. Moreover,
we show herein that 3 produces, to the best of our knowledge,
the best cell performance to date for a Ru-based dye with a
Co-based electrolyte at 1 sun.
The title sensitizer 3 was isolated according to the
synthetic approach developed for 1–2 (Scheme S1).[20] Briefly,
the cyclometalating ligand, ppy-(CF3)2, which was synthesized
by a Negishi coupling of 2-pyridylzinc bromide and 1-bromo-
2,4-bis(trifluoromethyl)benzene, [Ru(C6H6)Cl2]2, NaOH, and
KPF6 were reacted to furnish [Ru(CH3CN)4(ppy-(CF3)2)]PF6.
This complex was then treated with deeb (4,4’-diethylester-
2,2’-bipyridine) and dthbpy (2,2’-bis(5-hexylthiophen-2-yl)-
2,2’-bipyridine) to give the ethylester intermediate [Ru-
(deeb)(dthbpy)(ppy-(CF3)2)]PF6, which affords 3 in high
purity upon saponification.
1.84 eV for 3 extracted from the intersection of the absorption
and emission profiles[23] enables a determination of the
excited-state oxidation potential E(S+/S*) to be À0.85 V—a
value that is appropriately positioned for electron injection
into the TiO2 conduction band (ECB ꢀ À0.5 V vs. NHE).[24]
Photovoltaic data was recorded on DSSCs containing 3
with electrolytes relying on either the IÀ/I3À (denoted EL1) or
CoIII/CoII (denoted EL2) redox couples (Table 1). Data for
Table 1: Photovoltaic data obtained under AM1.5 irradiation.[a]
Entry Dye El[b] Light Level[c] Voc [V] Jsc [mAcmÀ2
]
FF
h [%]
1
2
3
4
1
EL1
1
1
1
0.63
0.65
0.66
0.64
0.66
0.73
0.72
0.78
0.49
9.7
13.3
16.3
8.6
0.66 4.1
0.73 6.3
0.68 7.3
0.74 8.3
0.68 8.8
0.66 6.9
0.64 5.5
0.61 6.2
0.51 0.96
N3 EL1
3
3
3
3
3
3
EL1
EL1 0.5
EL1 0.5
EL1
EL2
EL2
5[d]
6[e]
7
9.6
The UV/Vis spectrum of 3 reveals a broader and more
intense absorption envelope relative to both 1 and N3
(Figure 1). The absorption spectra of the dyes tethered to
1
1
1
1
14.3
11.8
13.1
3.8
8[d]
9
N3 EL2
[a] DSSC substrates consist of a 12 mm TiO2 active layer and 3 mm TiO2
scattering layer unless otherwise specified. [b] El=electrolyte; EL1: BMII
(1-butyl-3-methylimidazolium iodide) (0.6m), I2 (0.06m), NaI (0.1m),
tBP (4-tert-butylpyridine) (0.5m), GuSCN (guanidinium thiocyanate)
(0.1m) in MeCN; EL2: [Co(bpy)3](PF6)2 (0.21m), [Co(bpy)3](PF6)3
(0.033m), NaClO4 (0.1m), tBP (0.2m) in MeCN. [c] Light intensity
measured in suns. [d] Data recorded on same cell as previous entry
without a mask. [e] Substrate consists of a 6 mm TiO2 active layer and
3 mm TiO2 scattering layer.
Figure 1. UV/Vis spectra for MeOH solutions of N3, 1, and 3; the
emission spectrum of 3 (lex =555 nm) is also provided.
devices containing 1 and N3 are also provided as a benchmark
(Entries 1–2, Table 1). While our devices with EL1 reach h =
6.3% for N3, a markedly lower value is obtained for 1, which
TiO2 follow similar trends, but display a less pronounced
difference in intensities, presumably due to the lower surface
coverage of 3 (Figure S1). Intense p–p* transitions dominate
the spectrum of 3 below 350 nm, while metal-to-ligand
charge-transfer transitions comprise the broad bands cen-
tered at 404 nm (e = 2.9 ꢀ 104 mÀ1 cmÀ1) and 555 nm (e = 2.6 ꢀ
104 mÀ1 cmÀ1). These assignments were corroborated by DFT,
which predicts the two lowest-energy bands to contain
transitions from a metal-based HOMO-1 level to lowest
unoccupied molecular orbital (LUMO) and LUMO + 1 levels
that have orbital character localized to the polypyridyl ligands
(Figure S2). The shoulder at approximately 660 nm for 3 is
predicted by DFT to be a HOMO to LUMO transition
characterized by a low oscillator strength arising from the
orthogonality of the constituent orbitals; however, this band
may also be due to the direct population of the 3MLCT
state.[22] The intensities of the low-energy absorption bands of
3 are relatively higher than those of 1 and N3 due to the
presence of the thiophene units. Importantly, light absorption
by 3 occurs beyond 700 nm even with the electron-with-
is presumably due to an oxidation potential (i.e., + 0.70 V[20]
)
that is inappropriate for efficient dye regeneration. The
higher oxidation potential (and slightly enhanced absorb-
ance) of 3, however, leads to a stark improvement in cell
performance among the series (Entry 3, Table 1; Figure 2).
Indeed, efficiencies approaching 9% were obtained for cells
measured at lower light intensities and/or without a mask
(Entries 4–5, Table 1).
The superior cell performance of 3 is manifested in the
higher short-circuit photocurrent density (Jsc; 16.3 mAcmÀ2),
which arises from the thiophene groups enhancing the
À
drawing CF3 groups present.
The cyclic voltammogram of 3 in DMF reveals a reversible
one-electron oxidation at + 0.99 V vs. NHE and a reversible
ligand-based reduction wave at À1.17 V vs. NHE (Figure S3).
(For comparison, the first oxidation potentials for 1 and 2
occur at + 0.70 and + 0.93 V, respectively.) The E0À0 energy of
Figure 2. a) Current–voltage curves recorded under AM1.5 conditions
&
and b) IPCE data for DSSCs sensitized with 1 and 3. Squares ( ) and
*
circles ( ) represent cells measured with EL1 and EL2, respectively
(TiO2: 12 mm active and 3 mm scattering layers).
Angew. Chem. Int. Ed. 2011, 50, 10682 –10685
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 10683

Communications
absorbance profile and the alkyl groups suppressing recombi-
7.3% was achieved at AM1.5 and 8.3% at half that light
intensity. We also demonstrate that a noncorrosive cobalt
electrolyte achieves an efficiency of 5.5% and a photocurrent
of 11.8 mAcmÀ2 under full sun illumination. An examination
of the long-term stability of 3 in the DSSC is underway. These
results provide an important breakthrough for making high-
performance cyclometalated Ru dyes for use in the DSSC.
nation. This claim is corroborated by: 1) normalized incident-
photon-to-current efficiency (IPCE) measurements, which
display an enhanced external quantum efficiency between
600–700 nm for 3 (Figure S4); and 2) Nyquist plots of cells
measured in the dark at the voltage corresponding to their Voc
(Figure S5). (A transmission line model (as depicted in
Figure S6) was applied to the electrochemical impedance
spectroscopy (EIS) data; the results are summarized in
Table S1.[25]) For conventional cells (i.e., EL1 and a 12 mm
TiO2 active layer), the electron-transport resistance in TiO2
was found to be highest for 1 (64 W) and lowest for 3 (11 W).
Moreover, cells containing 3 had the highest resistance (30 W)
to electron recombination with EL1 and thus the longest
electron-diffusion length, Ln.
Given the superior absorbance of 3, we investigated the
performance of cells using thinner TiO2 films (Entry 6,
Table 1). Despite a two-fold reduction in the film thickness
of the substrate, the Jsc was lowered by only 12% to
14.3 mAcmÀ2. This diminution in current was countered by
an 11% increase in Voc to 0.73 V, which offsets the decrease in
photocurrent to afford a reasonably high h value of 6.9%. EIS
data indicates that the increase in Voc emanates from lower
recombination arising from improved charge collection (e.g.,
a low transport resistance of 2 W coupled to a larger charge-
transfer resistance with the electrolyte of 33 W), thereby
resulting in a cell that exhibits minor losses in performance
compared to the analogous cells with thicker substrates.
Recent reports have highlighted electrolytes that rely on
the CoIII/CoII redox shuttle and that can produce high h values
for organic dyes;[26–28] however, a similar performance rating
with Ru-based dyes at 1 sun has not yet been demonstrated;
for example, 3.9% using the [Co(dbbip)2]2+/3+ redox couple
Received: June 21, 2011
Published online: August 30, 2011
Keywords: cobalt · donor–acceptor systems ·
.
energy conversion · ruthenium · sensitizers
[1] B. OꢁRegan, M. Grꢂtzel, Nature 1991, 353, 737 – 740.
[2] A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo, H. Pettersson, Chem.
Rev. 2010, 110, 6595 – 6663.
[3] D. Shi, N. Pootrakulchote, R. Li, J. Guo, Y. Wang, S. M.
Zakeeruddin, M. Grꢂtzel, P. Wang, J. Phys. Chem. C 2008, 112,
17046 – 17050.
[4] L. Wang, X. Fang, Z. Zhang, Renewable Sustainable Energy Rev.
2010, 14, 3178 – 3184.
[5] P. T. Nguyen, A. R. Andersen, E. M. Skou, T. Lund, Sol. Energy
Mater. Sol. Cells 2010, 94, 1582 – 1590.
[6] P. T. Nguyen, B. X. T. Lam, A. R. Andersen, P. E. Hansen, T.
Lund, Eur. J. Inorg. Chem. 2011, 2533 – 2539.
[7] T. Bessho, S. Zakeeruddin, C. Y. Yeh, E. G. Diau, M. Grꢂtzel,
Angew. Chem. 2010, 122, 6796 – 6799; Angew. Chem. Int. Ed.
2010, 49, 6646 – 6649.
[8] E. Galoppini, Coord. Chem. Rev. 2004, 248, 1283 – 1297.
[9] P. Wang, C. Klein, R. Humphry-Baker, S. M. Zakeeruddin, M.
Grꢂtzel, J. Am. Chem. Soc. 2005, 127, 808 – 809.
[10] P. G. Bomben, K. C. D. Robson, P. A. Sedach, C. P. Berlinguette,
Inorg. Chem. 2009, 48, 9631 – 9643.
[11] S. H. Wadman, J. M. Kroon, K. Bakker, M. Lutz, A. L. Spek,
G. P. M. van Klink, G. van Koten, Chem. Commun. 2007, 1907 –
1909.
(dbbip = 2,6-bis(1’-butylbenzimidazol-2’-yl)pyridine)
with
Z907.[29] Because the electronic geometry of the HOMO for
3 differs from that of dyes bearing NCSÀ ligands (e.g., N3,
Z907), we were curious if the interaction with an electrolyte
derived from [Co(bpy)3]2+/3+ (e.g., EL2) could achieve a
higher cell performance. This line of inquiry was rewarded by
the observation that a DSSC containing 3 reached h = 5.5%
(Entries 7 and 8), which appears to be the highest value
obtained for any Ru-based dye with a Co-based electrolyte at
1 sun. To date, the highest efficiency for a Ru-based dye using
[Co(bpy)3]2+/3+ is 1.3%;[30] we were only able to obtain a cell
efficiency of 0.96% with N3 (Entry 9, Table 1). Note that the
high h and IPCE values (e.g., 50% over 400–600 nm) for 3
were obtained without the use of blocking layers.[30,31] We
observe a much larger Warburg feature in our EIS measure-
ments for EL2 than for EL1 (Figure S5), which we ascribe to
mass-transport limitations;[32] thus, investigations are under-
way to optimize cells of 3 with cobalt electrolytes using
thinner TiO2 films.
[12] K. C. D. Robson, B. Sporinova, B. D. Koivisto, T. Baumgartner,
A. Yella, Nazeeruddin, M. K., M. Grꢂtzel, C. P. Berlinguette,
Inorg. Chem. 2011, 50, 5494 – 5508.
[13] C.-C. Chou, K.-L. Wu, Y. Chi, W.-P. Hu, S. J. Yu, G.-H. Lee, C.-L.
Lin, P.-T. Chou, Angew. Chem. 2011, 123, 2102 – 2106; Angew.
Chem. Int. Ed. 2011, 50, 2054 – 2058.
[14] T. Bessho, E. Yoneda, J.-H. Yum, M. Guglielmi, I. Tavernelli, H.
Imai, U. Rothlisberger, M. K. Nazeeruddin, M. Grꢂtzel, J. Am.
Chem. Soc. 2009, 131, 5930 – 5934.
[15] P. G. Bomben, B. D. Koivisto, C. P. Berlinguette, Inorg. Chem.
2010, 49, 4960 – 4971.
[16] S. H. Wadman, M. Lutz, D. M. Tooke, A. L. Spek, F. Hartl,
R. W. A. Havenith, G. P. M. van Klink, G. van Koten, Inorg.
Chem. 2009, 48, 1887 – 1900.
[17] B. D. Koivisto, K. C. D. Robson, C. P. Berlinguette, Inorg. Chem.
2009, 48, 9644 – 9652.
[18] C.-J. Yao, L.-Z. Sui, H.-Y. Xie, W.-J. Xiao, Y.-W. Zhong, J. Yao,
Inorg. Chem. 2010, 49, 8347 – 8350.
[19] S. Ardo, G. J. Meyer, Chem. Soc. Rev. 2009, 38, 115 – 164.
[20] P. G. Bomben, K. D. Thꢃriault, C. P. Berlinguette, Eur. J. Inorg.
Chem. 2011, 1806 – 1814.
This study discloses the first high-efficiency trisheterolep-
tic cyclometalated Ru sensitizer. The efficacy of this NCSÀ-
free dye is a consequence of a sufficiently high RuIII/RuII
redox potential achieved by installing electron-withdrawing
groups at R2, thereby enhancing light absorption by placing
thiophenes at R1, and suppressing charge recombination with
terminal alkyl groups on said thiophenes. A performance of
[21] All redox potentials are reported vs
electrode (NHE) in this study.
a normal hydrogen
[22] N. Hirata, J. J. Lagref, E. J. Palomares, J. R. Durrant, M. K.
Nazeeruddin, M. Gratzel, D. Di Censo, Chem. Eur. J. 2004, 10,
595 – 602.
[23] Excitation of the lowest-energy peak of 3 produces a weak
emission band centred at 767 nm; Figure 1.
10684 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 10682 –10685

[24] M. Grꢂtzel, Nature 2001, 414, 338 – 344.
[29] H. Nusbaumer, S. M. Zakeeruddin, J.-E. Moser, M. Grꢂtzel,
Chem. Eur. J. 2003, 9, 3756 – 3763.
[30] B. M. Klahr, T. W. Hamann, J. Phys. Chem. C 2009, 113, 14040 –
14045.
[25] F. Fabregat-Santiago, J. Bisquert, G. Garcia-Belmonte, G.
Boschloo, A. Hagfeldt, Sol. Energy Mater. Sol. Cells 2005, 87,
117 – 131.
[26] S. M. Feldt, E. A. Gibson, E. Gabrielsson, L. Sun, G. Boschloo,
A. Hagfeldt, J. Am. Chem. Soc. 2010, 132, 16714 – 16724.
[27] D. Zhou, Q. Yu, N. Cai, Y. Bai, Y. Wang, P. Wang, Energy
Environ. Sci. 2011, 4, 2030 – 2034.
[31] J. W. Ondersma, T. W. Hamann, J. Phys. Chem. C 2010, 114,
638 – 645.
[32] J. J. Nelson, T. J. Amick, C. M. Elliott, J. Phys. Chem. C 2008,
112, 18255 – 18263.
[28] H. N. Tsao, C. Yi, T. Moehl, J.-H. Yum, S. M. Zakeeruddin,
M. K. Nazeeruddin, M. Grꢂtzel, ChemSusChem 2011, 4, 591 –
594.
Angew. Chem. Int. Ed. 2011, 50, 10682 –10685
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 10685