Spin-selective charge transport pathways through p-oligophenylene-linked donor-bridge-acceptor molecules

Article abstract of DOI:10.1021/ja907625k

A series of donor-bridge-acceptor (D-B-A) triads have been synthesized in which the donor, 3,5-dimethyl-4-(9-anthracenyl)julolidine (DMJ-An), and the acceptor, naphthalene-1,8:4,5-bis(dicarboximide) (NI), are linked by p-oligophenylene (Ph_n) br

Full text of DOI:10.1021/ja907625k

Published on Web 11/11/2009
Spin-Selective Charge Transport Pathways through
p-Oligophenylene-Linked Donor-Bridge-Acceptor Molecules
Amy M. Scott, Tomoaki Miura, Annie Butler Ricks, Zachary E. X. Dance,
Emilie M. Giacobbe, Michael T. Colvin, and Michael R. Wasielewski*
Department of Chemistry and Argonne-Northwestern Solar Energy Research (ANSER) Center,
Northwestern UniVersity, EVanston, Illinois 60208-3113
Received September 8, 2009; E-mail: m-wasielewski@northwestern.edu
Abstract: A series of donor-bridge-acceptor (D-B-A) triads have been synthesized in which the donor,
3,5-dimethyl-4-(9-anthracenyl)julolidine (DMJ-An), and the acceptor, naphthalene-1,8:4,5-bis(dicarboximide)
(NI), are linked by p-oligophenylene (Phn) bridging units (n ) 1-5). Photoexcitation of DMJ-An produces
DMJ+•-An-• quantitatively, so that An-• acts as a high potential electron donor, which rapidly transfers an
electron to NI yielding a long-lived spin-coherent radical ion pair (DMJ+•-An-Phn-NI-•). The charge transfer
properties of 1-5 have been studied using transient absorption spectroscopy, magnetic field effects (MFEs)
on radical pair and triplet yields, and time-resolved electron paramagnetic resonance (TREPR) spectroscopy.
The charge separation (CS) and recombination (CR) reactions exhibit exponential distance dependencies
with damping coefficients of ꢀ ) 0.35 Å-1 and 0.34 Å-1, respectively. Based on these data, a change in
mechanism from superexchange to hopping was not observed for either process in this system. However,
the CR reaction is spin-selective and produces the singlet ground state and both 3*An and 3*NI. A kinetic
analysis of the MFE data shows that superexchange dominates both pathways with ꢀ ) 0.48 Å-1 for the
singlet CR pathway and ꢀ ) 0.35 Å-1 for the triplet CR pathway. MFEs and TREPR experiments were
used to measure the spin-spin exchange interaction, 2J, which is directly related to the electronic coupling
matrix element for CR, V2CR. The magnitude of 2J also shows an exponential distance dependence with a
damping coefficient R ) 0.36 Å-1, which agrees with the ꢀ values obtained from the distance dependence
for triplet CR. These results were analyzed in terms of the bridge molecular orbitals that participate in the
charge transport mechanism.
Introduction
has inspired the design of organic donor-bridge-acceptor (D-
B-A) systems that exhibit efficient, controlled energy and charge
Charge transport plays an integral role in many fields
including physics, chemistry, and biology and has revolutionized
how simple molecular systems with unique photochemical and
electronic properties can be utilized as photofunctional materials
for solar energy conversion,1 photocatalysis,2,3 and optoelec-
tronic nanodevices.4,5 In particular, there is a great need to
develop new, tunable materials that efficiently use light to
separate charge and demonstrate distance-independent charge
transport. The multistep electron transfer pathway in photosyn-
thetic reaction centers converts photon energy into chemical
energy with near unity efficiency using highly ordered as-
semblies of chromophores to separate charge over distances of
35 Å within these membrane-bound proteins.6,7 The photoge-
neration of radical ion pairs (RPs) within reaction center proteins
transport as a means of capturing solar energy.8 Charge transport
has been studied in covalently linked D-B-A systems with
modified bridge molecules, such as proteins,9 DNA,10-12
chromophores,13 saturated14 and unsaturated phenylene spac-
ers,15-19 but, often, multiple charge transport pathways exist
and the factors that influence them still remain poorly under-
stood. There is a growing effort to understand how the electronic
structure and composition of the bridge governs the charge
transport mechanism and thus influences the lifetimes of
photogenerated RPs (D+•-B-A-•).20-25
The utilization of rigid donors and acceptors with varying
bridge lengths allows for the measurement of distance dependent
electron transfer rates with the aim of elucidating the superex-
change and hopping charge transfer regimes within molecular
(1) Baik, C.; Kim, D.; Kang, M. S.; Kang, S. O.; Ko, J.; Nazeeruddin,
M. K.; Gratzel, M. J. Photochem. Photobiol., A 2009, 201 (2-3),
168–174.
(2) Kunkely, H.; Vogler, A. Angew. Chem., Int. Ed. 2009, 48 (9), 1685–
1687.
(8) Wasielewski, M. R. J. Org. Chem. 2006, 71 (14), 5051–5066.
(9) Gray, H. B.; Winkler, J. R. Proc. Natl. Acad. Sci. U.S.A. 2005, 102
(10), 3534–3539.
(3) Barber, J. Chem. Soc. ReV. 2009, 38 (1), 185–196.
(4) Straight, S. D.; Andreasson, J.; Kodis, G.; Bandyopadhyay, S.;
Mitchell, R. H.; Moore, T. A.; Moore, A. L.; Gust, D. J. Am. Chem.
Soc. 2005, 127 (26), 9403–9409.
(10) Jortner, J.; Bixon, M.; Langenbacher, T.; Michel-Beyerle, M. E. Proc.
Natl. Acad. Sci. U.S.A. 1998, 95, 12759–12765.
(11) Lewis, F. D.; Wu, T.; Zhang, Y.; Letsinger, R. L.; Greenfield, S. R.;
Wasielewski, M. R. Science 1997, 277 (5326), 673–676.
(12) Daublain, P.; Siegmund, K.; Hariharan, M.; Vura-Weis, J.; Wasielewski,
M. R.; Lewis, F. D.; Shafirovich, V.; Wang, Q.; Raytchev, M.; Fiebig,
T. Photochem. Photobiol. Sci. 2008, 7 (12), 1501–1508.
(5) Flood, A. H.; Stoddart, J. F.; Steuerman, D. W.; Heath, J. R. Science
2004, 306 (5704), 2055–2056.
(6) Wasielewski, M. R. Chem. ReV. 1992, 92, 435–461.
(7) Parson, W. W. Science 2007, 316 (5830), 1438–1439.
9
10.1021/ja907625k CCC: $40.75  2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 17655–17666 17655

A R T I C L E S
Scott et al.
systems.10,26-29 Superexchange is the virtual mediation of
charge transport from donor to acceptor by bridge orbitals
energetically well-separated from those of the donor and
acceptor. Charge transfer occurs in a single step with a rate
that decays exponentially with distance that is described by the
damping factor ꢀ. In contrast, sequential charge transport or
hopping occurs from the donor to the bridge, as well as between
bridge sites. Ultimately, charge trapping occurs at the acceptor
whose electron affinity is greater than that of the bridge. Charge
hopping by direct injection of charge from the donor to the
bridge, and/or between bridge sites, can occur by tunneling
through small potential barriers between successive sites,30-32
or by thermal activation, where the energy barrier between sites
is lowered by internal or solvent nuclear reorganization to allow
the system to surmount the barriers.30 Charge transport rates
for the hopping mechanism are only weakly distance dependent
(1/r) allowing charge to move efficiently at long distances, so
that systems exhibiting this mechanism are termed “molecular
wires”.33-35 Although molecular systems may display a com-
bination of charge transport mechanisms, there are certain
criteria that drive photoinduced charge transport in D-B-A
systems toward a particular mechanism, such as (1) the
magnitude of the donor-to-bridge charge injection barrier as
characterized by the energy levels of the system,36 (2) the energy
required for internal and solvent nuclear reorganization, and (3)
Figure 1. Charge transfer scheme for 1-5.
(13) Kobori, Y.; Yamauchi, S.; Akiyama, K.; Tero-Kubota, S.; Imahori,
H.; Fukuzumi, S.; Norris, J. R., Jr. Proc. Natl. Acad. Sci. U.S.A. 2005,
102 (29), 10017–10022.
Figure 2. Schematic of radical ion pair energy levels as a function of
magnetic field (2J > 0).
(14) Paulson, B. P.; Miller, J. R.; Gan, W. X.; Closs, G. J. Am. Chem.
Soc. 2005, 127 (13), 4860–4868.
the electronic coupling of the bridge to the donor and acceptor.
The overall electronic coupling matrix element, VDA, describes the
interaction between the donor and acceptor and is important for
determining rates of nonadiabatic charge transfer reactions.37-40
(15) Lembo, A.; Tagliatesta, P.; Guldi, D. M.; Wielopolski, M.; Nuccetelli,
M. J. Phys. Chem. A 2009, 113 (9), 1779–1793.
(16) Albinsson, B.; Mårtensson, J. J. Photochem. Photobiol., C 2008, 9
(3), 138–155.
(17) Wielopolski, M.; Atienza, C.; Clark, T.; Guldi, D. M.; Martin, N.
Chem.sEur. J. 2008, 14 (21), 6379–6390.
For charge recombination (CR), a fourth criterion must be
added, that of spin selectivity. The photogenerated singlet RP
may undergo electron-nuclear hyperfine coupling-induced
radical pair intersystem crossing (RP-ISC) to produce the triplet
(18) Martin, N.; Giacalone, F.; Segura, J. L.; Martin, N.; Guldi, D. M.
Synth. Met. 2004, 147 (1-3), 57–61.
(19) Guldi, D. M.; Luo, C.; Swartz, A.; Gomez, R.; Segura, J. L.; Martin,
N. J. Phys. Chem. A 2004, 108 (3), 455–467.
(20) Albinsson, B.; Eng, M. P.; Pettersson, K.; Winters, M. U. Phys. Chem.
Chem. Phys. 2007, 9 (44), 5847–5864.
1
3
RP, (D+•-B-A-•) f (D+•-B-A-•). The subsequent CR is spin
selective; i.e., the singlet RP recombines to the singlet ground
state, and the triplet RP recombines to yield the neutral local
triplet (Figure 1). Application of a static magnetic field causes
Zeeman splitting of the RP triplet sublevels, and varying the
field strength modulates the efficiency of RP-ISC by adjusting
the energies of triplet sublevels relative to that of the singlet
level (Figure 2). When the Zeeman splitting of the triplet RP
sublevels equals the intrinsic singlet-triplet splitting, 2J, of the
RP, there is an increase in the RP-ISC rate. This increase
translates into a maximum in triplet RP production and therefore
a maximum in neutral local triplet yield upon CR. By monitoring
the yield of local triplet production as a function of applied
magnetic field, the magnitude of 2J can be measured directly41,42
(21) Goldsmith, R. H.; Sinks, L. E.; Kelley, R. F.; Betzen, L. J.; Liu, W.;
Weiss, E. A.; Ratner, M. A.; Wasielewski, M. R. Proc. Natl. Acad.
Sci. U.S.A. 2005, 102 (10), 3540–3545.
(22) Goldsmith, R. H.; Vura-Weis, J.; Scott, A. M.; Borkar, S.; Sen, A.;
Ratner, M. A.; Wasielewski, M. R. J. Am. Chem. Soc. 2008, 130
(24), 7659–7669.
(23) Bixon, M.; Jortner, J. J. Chem. Phys. 1997, 107 (13), 5154–5170.
(24) McConnell, H. M. J. Chem. Phys. 1961, 35 (2), 508–515.
(25) Paddon-Row, M. N.; Oliver, A. M.; Warman, J. M.; Smit, K. J.; de
Haas, M. P.; Oevering, H.; Verhoeven, J. W. J. Phys. Chem. 1988,
92 (24), 6958–6962.
(26) Lambert, C.; Noll, G.; Schelter, J. Nat. Mater. 2002, 1 (1), 69–73.
(27) Davis, W. B.; Wasielewski, M. R.; Ratner, M. A.; Mujica, V.; Nitzan,
A. J. Phys. Chem. 1997, 101, 6158–6164.
(28) Johnson, M. D.; Miller, J. R.; Green, N. S.; Closs, G. L. J. Phys.
Chem. 1989, 93 (4), 1173–6.
(29) Bixon, M.; Jortner, J. AdV. Chem. Phys. 1999, 106, 3.
(30) Berlin, Y. A.; Burin, A. L.; Ratner, M. A. Chem. Phys. 2002, 275,
61–74.
(31) Berlin, Y. A.; Burin, A. L.; Ratner, M. A. J. Phys. Chem. A 2000,
104, 443–445.
(37) Barbara, P. F.; Meyer, T. J.; Ratner, M. A. J. Phys. Chem. 1996,
100 (31), 13148–13168.
(32) Berlin, Y. A.; Hutchison, G. R.; Rempala, P.; Ratner, M. A.; Michl,
J. J. Phys. Chem. A 2003, 107 (19), 3970–3980.
(38) Cave, R. J.; Newton, M. D. Chem. Phys. Lett. 1996, 249 (1,2), 15–
19.
(33) Davis, W. B.; Svec, W. A.; Ratner, M. A.; Wasielewski, M. R. Nature
1998, 396, 60–63.
(39) Cave, R. J.; Newton, M. D.; Kumar, K.; Zimmt, M. B. J. Phys. Chem.
1995, 99, 17501–17504.
(34) Weiss, E. A.; Ahrens, M. J.; Sinks, L. E.; Gusev, A. V.; Ratner,
M. A.; Wasielewski, M. R. J. Am. Chem. Soc. 2004, 126, 5577–
5584.
(40) Roest, M. R.; Oliver, A. M.; PaddonRow, M. N.; Verhoeven, J. W.
J. Phys. Chem. A 1997, 101 (27), 4867–4871.
(41) Anderson, P. W. Phys. ReV. 1959, 115 (1), 2–13.
(42) Shultz, D. A.; Fico, R. M., Jr.; Bodnar, S. H.; Kumar, R. K.;
Vostrikova, K. E.; Kampf, J. W.; Boyle, P. D. J. Am. Chem. Soc.
2003, 125 (38), 11761–11771.
(35) Joachim, C.; Vinuesa, J. F. Europhys. Lett. 1996, 33 (8), 635–40.
(36) Samanta, M. P.; Tian, W.; Datta, S.; Henderson, J. I.; Kubiak, C. P.
Phys. ReV. B: Condens. Matter 1996, 53 (12), R7626–R7629.
9
17656 J. AM. CHEM. SOC. VOL. 131, NO. 48, 2009

Spin-Selective Charge Transport Pathways
A R T I C L E S
and can be used to estimate VDA because 2J ∝ V2DA 34,43-45 The
.
mechanistic details of the radical pair intersystem crossing
mechanism (RP-ISC) and the theory behind the magnetic field
effects (MFEs) have been researched extensively46-49 and
applied to many donor-acceptor systems34,50-57 including
biological58-62 systems.
Time-resolved EPR (TREPR) measurements yield additional
information on the spin dynamics of the RPs. If 2J and the
spin-spin dipolar interaction, D, of the RP are small, S-T0
mixing within D+•-B-A-• will result in the formation of a spin-
correlated RP (SCRP) that can be recognized by the electron
spin polarization pattern of its EPR transitions, i.e., enhanced
absorption (A) or emission (E).63-65 The TREPR spectrum of
a SCRP consists of a pair of antiphase doublets centered at the
g factors of each radical, where the doublet splitting is
determined principally by 2J and D, and these four lines are
split further by the hyperfine couplings of each radical.
Simulation of these spectra yield information critical to
understanding the spin dynamics of RPs photogenerated
within D-B-A systems, as well as determining the magnitude
of 2J for systems in which its value is too small to measure
accurately by MFE measurements.
Figure 3. D-B-A molecules used in this study.
(DMJ+•-An-Phn-NI-•). Photoexcitation of DMJ-An produces
DMJ+•-An-• quantitatively, so that An-• acts as a high potential
electron donor. Using TREPR spectroscopy at 85 K, we
observed a preference for hole transfer over electron transfer66
for RP recombination to selectively form (DMJ-An-Phn-3*NI)
over (DMJ-3*An-Phn-NI) for 1-4. The work described here
builds significantly upon this initial study and focuses on (1)
determining how VDA depends on molecular structure by
analyzing the spin-spin exchange interaction (2J) between the
unpaired spins of the RP as determined by MFEs and TREPR
and (2) establishing that the distance dependence of CR in
D-B-A molecules is spin-selective using a kinetic model to
separate the spin-selective CR rates, kCRS and kCRT, by analyzing
MFEs on the observed RP populations as well as the neutral
triplet state populations that result from CR.
We recently reported the results of our investigations into
photoinduced charge transport in a D-B-A molecular system,
in which the donor, 3,5-dimethyl-4-(9-anthracenyl)julolidine
(DMJ-An), and the acceptor, naphthalene-1,8:4,5-bis(dicarbox-
imide) (NI), are linked by p-oligophenylene (Phn) bridging units
(1-4) (Figure 3) to produce a long-lived spin-coherent RP
Experimental Section
The synthesis and characterization of compounds 1-4 have been
reported,66 and compound 5 can be found in the Supporting
Information. Samples for nanosecond transient absorption spec-
troscopy were placed in a 10 mm path length quartz cuvette and
freeze-pump-thawed five times. The samples were excited with
7 ns, 2.5 mJ, 416 nm using the frequency-tripled output of a
Continuum Precision II 8000 Nd:YAG laser pumping a Continuum
Panther OPO. The transient absorption spectrum of 3*NI was
obtained from a deoxygenated (freeze-pump-thaw) solution of
50:50 toluene:ethyl iodide using 355 nm irradiation.67 The transient
absorption spectra of 3*An and DMJ-3*An were obtained from
deoxygenated (freeze-pump-thaw) toluene solutions using 355
nm irradiation. The excitation pulse was collimated to a 5 mm
diameter spot and matched to the diameter of the probe pulse
generated using a xenon flashlamp (EG&G Electro-Optics FX-200).
Kinetic traces were observed from 430-800 nm every 5 nm using
a monochromator and photomultiplier tube with high voltage
applied to only four dynodes (Hamamatsu R928) and recorded with
a LeCroy Wavesurfer 42Xs oscilloscope interfaced to a customized
Labview program (Labview v. 8.5.2). The total instrument response
time is 7 ns and is determined primarily by the laser pulse duration.
Analysis of the kinetic data was performed at multiple wavelengths
using a Levenberg-Marquardt nonlinear least-squares fit to a
general sum-of-exponentials function with an added Gaussian
function to account for the finite instrument response. For the
magnetic field effect experiments, the sample cuvette was placed
between the poles of a Walker Scientific HV-4W electromagnet
powered by a Walker Magnion HS-735 power supply and the field
strength was measured by a Lakeshore gaussmeter with a Hall effect
probe. The electromagnet and gaussmeter were interfaced with
Labview, allowing measurements and control of the magnetic field
to (1 × 10-5 T during the data acquisition. To maintain sample
integrity during the experiment, a probe light shutter was used to
(43) Kobori, Y.; Sekiguchi, S.; Akiyama, K.; Tero-Kubota, S. J. Phys.
Chem. A 1999, 103, 5416–5424.
(44) Paddon-Row, M. N.; Shephard, M. J. J. Phys. Chem. A 2002, 106
(12), 2935–2944.
(45) Volk, M.; Haberle, T.; Feick, R.; Ogrodnik, A.; Michel-Beyerle, M. E.
J. Phys. Chem. 1993, 97 (38), 9831–9836.
(46) Weller, A.; Staerk, H.; Treichel, R. Faraday Discuss. Chem. Soc.
1984, 78, 271–278.
(47) Hoff, A. J.; Gast, P.; van der Vos, R.; Franken, E. M.; Lous, E. J. Z.
Phys. Chem. 1993, 180, 175–192.
(48) Till, U.; Hore, P. J. Mol. Phys. 1997, 90 (2), 289–296.
(49) Steiner, U. E.; Ulrich, T. Chem. ReV. 1989, 89, 51–147.
(50) Schulten, K.; Staerk, H.; Weller, A.; Werner, H. J.; Nickel, B. Z.
Phys. Chem. 1976, 101 (1-6), 371–390.
(51) Tanimoto, Y.; Okada, N.; Itoh, M.; Iwai, K.; Sugioka, K.; Takemura,
F.; Nakagaki, R.; Nagakura, S. Chem. Phys. Lett. 1987, 136 (1), 42–
46.
(52) Sakaguchi, Y.; Hayashi, H. J. Phys. Chem. A 1997, 101, 549–555.
(53) Werner, U.; Kuhnle, W.; Staerk, H. J. Phys. Chem. 1993, 97, 9280–
9287.
(54) Weiss, E. A.; Ratner, M. A.; Wasielewski, M. R. J. Phys. Chem. A
2003, 107 (19), 3639–3647.
(55) Lukas, A. S.; Bushard, P. J.; Weiss, E. A.; Wasielewski, M. R. J. Am.
Chem. Soc. 2003, 125, 3921–3930.
(56) Tadjikov, B.; Smirnov, S. Phys. Chem. Chem. Phys. 2001, 3 (2),
204–212.
(57) Tsentalovich, Y. P.; Morozova, O. B.; Avdievich, N. I.; Ananchenko,
G. S.; Yurkovskaya, A. V.; Ball, J. D.; Forbes, M. D. E. J. Phys.
Chem. A 1997, 101 (47), 8809–8816.
(58) Blankenship, R. E.; Schaafsma, T. J.; Parson, W. W. Biochim.
Biophys. Acta 1977, 461, 297–305.
(59) Plato, M.; Mo¨bius, K.; Michel-Beyerle, M. E.; Bixon, M.; Jortner,
J. J. Am. Chem. Soc. 1988, 110 (22), 7279–85.
(60) van der Vos, R.; Hoff, A. J. Biochim. Biophys. Acta 1995, 1228 (1),
73–85.
(61) Werner, H.-J.; Schulten, K.; Weller, A. Biochim. Biophys. Acta 1978,
502, 255–268.
(62) Norris, J. R.; Bowman, M. K.; Budil, D. E.; Tang, J.; Wraight, C. A.;
Closs, G. L. Proc. Natl. Acad. Sci. U.S.A. 1982, 79, 5532–5536.
(63) Closs, G. L.; Forbes, M. D. E.; Norris, J. R. J. Phys. Chem. 1987,
91 (13), 3592–3599.
(66) Dance, Z. E. X.; Ahrens, M. J.; Vega, A. M.; Ricks, A. B.;
McCamant, D. W.; Ratner, M. A.; Wasielewski, M. R. J. Am. Chem.
Soc. 2008, 130 (3), 830–832.
(64) Hore, P. J.; Hunter, D. A.; McKie, C. D.; Hoff, A. J. Chem. Phys.
Lett. 1987, 137, 495–500.
(65) Levanon, H.; Hasharoni, K. Prog. React. Kinet. 1995, 20, 309–46.
(67) Green, S.; Fox, M. A. J. Phys. Chem. 1995, 99 (40), 14752–14757.
9
J. AM. CHEM. SOC. VOL. 131, NO. 48, 2009 17657

A R T I C L E S
Scott et al.
block the sample from irradiation when transient absorption kinetics
were not being collected. The triplet yield was monitored at 480
and 430 nm, and kinetic traces were collected in increments of
0.3, 1.5, or 5.0 mT with zero field ∆A(B ) 0) collection after four
or five steps. To compensate for possible sample degradation, zero
field kinetics were collected during the experiment in four- or five-
step increments and plotted and fit with polynomial or linear trend
lines. These functions were used to calculate the relative RP yield
or triplet yield as a function of applied field strength (B) and plotted
as ∆A(B)/∆A(B ) 0). The results presented are an average of three
or more experiments conducted on separate days with freshly
prepared samples in spectrophotometric or freshly distilled ACS-
grade toluene. Transient absorption experiments for 4 and 5 were
performed at varying concentrations, ranging from OD ) 0.30 to
0.05 using 416 nm excitation to ensure that the microsecond time
scale electron transfer kinetics monitored only intramolecular
processes, not intermolecular processes. The three-dimensional data
sets of ∆A vs time and λ were subjected to singular value
decomposition and global fitting to obtain the kinetic time constants
and their decay associated spectra using Surface Xplorer software.68
Femtosecond transient absorption measurements were made
either using the 420 nm frequency-doubled output from a regen-
eratively amplified titanium sapphire laser system operating at 2
kHz or the 400 nm frequency-doubled output from a regeneratively
amplified titanium sapphire laser system operating at 1 kHz as the
excitation pulse. A white light continuum probe pulse was generated
by focusing the IR fundamental into a 1-mm sapphire disk.
Detection with a CCD spectrograph has previously been de-
scribed.34 The samples were irradiated with 0.5-1.0 µJ per pulse
focused to a 200-µm spot. Samples were placed in a 2-mm path
length cuvette, and the total instrument response time for the
pump-probe experiment was 180 fs. Transient absorption kinetics
were fit to a sum of exponentials with a Gaussian instrument
function using Levenberg-Marquardt least-squares fitting. Fem-
tosecond transient absorption kinetic traces for 1-4 have been
reported.66
For EPR measurements at X-band, toluene solutions of 1-5
(∼10-4 M) were loaded into quartz tubes (4 mm o.d. × 2 mm i.d.),
subjected to five freeze-pump-thaw degassing cycles on a vacuum
line (10-4 mBar), and sealed using a hydrogen torch. TREPR
measurements were made at X-band using continuous wave (CW)
microwaves and direct detection on a Bruker E-580 spectrometer.
Samples were photoexcited at 416 nm using the output from a
frequency tripled, H2-Raman shifted Nd:YAG laser (1-2 mJ/pulse,
7 ns, 10 Hz, QuantaRay DCR-2). The polarization of the laser was
set to 54.7° relative to the direction of the static magnetic field to
avoid magnetophotoselection effects on the spectra. Following
photoexcitation, kinetic traces of the transient magnetization were
accumulated under CW microwave irradiation (6-20 mW). The
field modulation was disabled to achieve a Q/πν ≈ 30 ns instrument
response function (IRF), where Q is the quality factor of the
resonator and ν is the resonant frequency, while microwave signals
in emission (e) and/or enhanced absorption (a) were detected in
both the real and the imaginary channels (quadrature detection).
Sweeping the magnetic field gave 2D spectra versus both time and
magnetic field. For each kinetic trace, the signal acquired prior to
the laser pulse was subtracted from the data. Kinetic traces recorded
at magnetic field values off-resonance were considered background
signals, whose average was subtracted from all kinetic traces. The
spectra were subsequently phased into a Lorentzian part and a
dispersive part, and the former, also known as the imaginary
magnetic susceptibility ꢁ”, is presented. Temperatures for
experiments were controlled by Oxford Instruments CF935 using
liquid N2.
Figure 4. Ground-state UV-vis spectrum of 1-3 in toluene at 295 K.
(CT) absorption maximum at 367 nm with a broad emission
maximum at 519 nm, resulting in an excited singlet CT state
energy in toluene of 2.89 eV.66,69 The steady-state absorption
spectra of DMJ-An-Phn-NI for n ) 1-3 are shown in Figure 4.
The absorption maxima at 355 and 325 nm are due to An and
NI vibronic bands with an underlying CT absorption band for
DMJ-An. The NI electron acceptor has vibronic absorption
bands at 343, 363, and 382 nm.70 The UV band near 300 nm
results from absorption of the phenyl bridging unit and increases
in intensity with longer bridge lengths. Steady-state fluorescence
measurements have been previously reported66 for 1-4 and
show quantum yields for emission ΦCT < 0.01 at room
temperature. Molecule 5 has a similarly low emission quantum
yield.
Energy Levels and Molecular Structures. Given that photo-
excitation of DMJ-An results in quantitative charge separation
to produce DMJ+•-An-• with a spectroscopically determined
energy of 2.89 eV in toluene,66 the energy levels for the charge
separated states, DMJ+•-An-Phn-NI-•, DMJ-An-Ph+•-NI-•, DMJ-
An+•-Phn-NI-•, and DMJ+•-An-Phn-•NI were determined in
toluene using eq 1
e2
εS rI
1
1
rF
∆GF ) ∆GI + sign(EI - EF) +
-
(1)
(
)
if EF > EI then sign ) (-), and if EI > EF then sign ) (+). ∆GI
and ∆GF are the energies above ground state for the initial and
final ion pairs, respectively, EI and EF are the redox potentials
for the initial and final ions, respectively, between which the
electron is transferred, rI and rF are the initial and final ion pair
distances, respectively, e is the electronic charge, and εS is the
static dielectric constant of the solvent (εS ) 2.38 for toluene).
The distances rI and rF between the donor, bridge, and acceptor
components were determined from the energy-minimized struc-
tures of DMJ-An-Phn-NI determined with density functional
theory (DFT) using Becke’s71 three-parameter hybrid functional
with Lee, Yang, and Parr72 correction functional (B3LYP) and
the STO-3G basis set. The redox potentials of DMJ, An, Phn,
and NI are given in the Supporting Information and the ion
pair distances and energies are listed in Table S1 and illustrated
in Figure 5.
The DFT calculations reveal that the HOMOs of 1-5 are
localized on DMJ and the LUMOs are confined to the NI
Results
(69) Herbich, J.; Kapturkiewicz, A. Chem. Phys. Lett. 1997, 273, 9–17.
(70) Greenfield, S. R.; Svec, W. A.; Gosztola, D.; Wasielewski, M. R.
J. Am. Chem. Soc. 1996, 118 (28), 6767–6777.
Steady-State Spectroscopy. The ground-state absorption spec-
trum of DMJ-An in toluene exhibits a broad charge transfer
(71) Becke, A. D. J. Chem. Phys. 1993, 98 (2), 1372–1377.
(72) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37 (2), 785–789.
(68) Surface Xplorer, Ultrafast Systems LLC, Sarasota, FL 34234.
9
17658 J. AM. CHEM. SOC. VOL. 131, NO. 48, 2009

Spin-Selective Charge Transport Pathways
A R T I C L E S
3*An-Phn-NI) (ET ) 1.85 eV).77 3*NI and 3*An exhibit broad
absorption features at 480 and 430 nm78 (Figure 6), respectively,
which persist on the microsecond time scale and appear as a
plateaus in the kinetic traces.
To help understand these transient absorption kinetics, the
absorption spectra of 3*NI and 3*An were compared to that of
DMJ-3*An, which forms in appreciable amounts upon CR of
1(DMJ+•-An-•).79 The 3*NI and 3*An absorption spectra were
normalized, added, and overlaid with the nanosecond transient
absorption spectrum of 3 acquired 1.5 µs after the excitation
pulse (Figure S2E). The spectra are qualitatively identical and
confirm that 3*An and 3*NI species both form at 295 K upon
CR (Figures 6 and S2A-C). This behavior contrasts with that
observed at 85 K, where 3*NI forms initially and is followed
by triplet-triplet energy transfer (TEnT) to yield 3*An.66 (see
Discussion). The kinetics at the 430 nm 3*An absorption exhibit
a plateau lasting into the microsecond regime. While the 3*An
absorption overlaps with those of NI-• and 3*NI, the extinction
coefficient of 3*An (ε ) 42 000 ( 4000 M-1 cm-1) at 430 nm
is considerably larger those of NI-• and 3*NI,80 so that the
analysis of the 430 nm kinetic traces monitor the concentration
of 3*An reasonably well.
The transient absorption spectra for 1 exhibit a weak 430
nm band, and the kinetics show a long-lived plateau with a rise
time of 7 ns, which is within the instrument response function.
The corresponding spectra for 2-4 show transient decays of
NI-• at 480 nm having lifetimes that match the appearance
kinetics for 3*An at 430 nm. For example, the 480 nm decay of
NI-• for 3 occurs with τ ) 345 ( 4 ns and is within
experimental uncertainty of the 340 ( 6 ns rise at 430 nm
attributed to 3*An (Figure S3). This confirms that 3*An is
generated upon the CR of DMJ+•-An-Phn-NI-•, as opposed to
being produced by CR from the initial ion pair state, (DMJ+•-
An-•). This is expected because the reaction DMJ+•-An-•-Phn-
NI f DMJ+•-An-Phn-NI-• occurs on a subnanosecond time scale
and is considerably faster than the 45 ns decay time of 1(DMJ+•-
An-•) that leads to about a 45% yield of DMJ-3*An.79 Further
evidence indicating that CR produces both 3*NI and 3*An is
derived from the decay-associated spectra (Figure 6B), which
reveal two major components with absorption features consistent
with the decay of NI-• and simultaneous rise of 3*NI and 3*An.
The species-associated time constants for NI-• decay are very
similar to those obtained by fitting the individual kinetics (Table
1). The spectra for 1-5 also show that the 3*An yield is greater
for 3 and 4 having the longer bridges relative to that for 1 and
2 having the shorter bridges. The transient absorption signal
for 5 at 430 nm is small in comparison to those of 1-4 because
the sample concentration is much lower, so that diffusion leading
to intermolecular CR does not contribute to the electron transfer
rate analysis.
Figure 5. Energy level diagram for 1-5 ion pair states based on eq 1.
acceptor. Energy-minimized structures of the ground state and
the corresponding radical ions of DMJ, DMJ-An, An-Phn, Phn,
and NI were calculated with B3LYP/6-31G*, and the HOMO
and LUMO energies are given in Figure S1. The internal
reorganization energies, λi, were calculated for CR within 1-5
by performing a single-point calculation (UHF) B3LYP/6-31G*
on the radical ions of the donor, bridge, and acceptors in the
DFT-optimized ground state and subtracting the self-consistent
field (SCF) energy of the relaxed ionic structure from that of
the unrelaxed ground-state structure (Table S2).73
Transient Absorption Spectroscopy. Selective photoexcitation
into the red edge of the CT absorption band of DMJ-An in
toluene at 416 nm results in nearly quantitative formation of
1(DMJ+•-An-•) with absorption peaks of DMJ+• at 480 nm and
An-• at 680 nm in toluene, which decay with τ ) 45 ns.66 The
charge separation (CS) reactions, determined by femtosecond
transient absorption, of DMJ+•-An-•-Phn-NI f DMJ+•-An-Phn-
NI-• for 1-5 in toluene are very rapid and nearly quantitative.
Nanosecond transient absorption was used to extract charge
recombination rates (CR) for 1-5 in toluene, and the rate
constants for these reactions were extracted from the kinetics
obtained at the prominent NI-• absorption bands at 480 and 610
70
nm (ε480 ) 30 000 M-1cm-1
)
and are given in Table 1. In
comparison, the DMJ+• absorption at 490 nm is weak (εmax
)
74
4500 M-1cm-1
)
and overlaps with that of NI-•. The nano-
second transient absorption spectra for 3 are shown in Figure
6A, and the results of singular value decomposition (SVD)
analysis with global fitting for 3 are shown in Figure 6B. The
spectra for 2, 4, and 5 are similar to those of 3 and given in
Figure S2. Following rapid CS, the initially formed singlet RP,
1(DMJ+•-An-Phn-NI-•), undergoes rapid intersystem crossing
induced by electron-nuclear hyperfine coupling within the
Magnetic Field Effect and TREPR Experiments. The MFE
experiments that directly measure the magnitude of the spin-spin
exchange interaction, 2J, were performed at room temperature
for 1-4 in toluene, and the magnetic field dependence of the
DMJ-An-Phn-3*NI yields formed by RP-ISC are shown for 1
3
radicals to produce the triplet RP, (DMJ+•-An-Phn-NI-•). The
energy level of the triplet RP state is above that of 3*An and
3*NI, so that spin-selective CR can occur from the triplet RP to
produce (DMJ-An-Phn-3*NI) (ET ) 2.03 eV)75,76 and (DMJ-
(73) Weiss, E. A.; Tauber, M. J.; Kelley, R. F.; Ahrens, M. J.; Ratner,
M. A.; Wasielewski, M. R. J. Am. Chem. Soc. 2005, 127 (33), 11842–
11850.
(77) Langelaar, J.; Rettschnick, R. P. H.; Hoijtink, G. J. J. Chem. Phys.
1971, 54 (1), 1–7.
(78) Livingston, R.; Tanner, D. W. Trans. Faraday Soc. 1958, 54 (6),
765–771.
(74) Okada, T.; Fujita, T.; Kubota, M.; Masaki, S.; Mataga, N.; Ide, R.;
Sakata, Y.; Misumi, S. Chem. Phys. Lett. 1972, 14 (5), 563–568.
(75) Wiederrecht, G. P.; Svec, W. A.; Wasielewski, M. R.; Galili, T.;
Levanon, H. J. Am. Chem. Soc. 2000, 122 (40), 9715–9722.
(76) Ganesan, P.; Baggerman, J.; Zhang, H.; Sudhoelter, E. J. R.; Zuilhof,
H. J. Phys. Chem. A 2007, 111 (28), 6151–6156.
(79) Dance, Z. E. X.; Mickley, S. M.; Wilson, T. M.; Ricks, A. B.; Scott,
A. M.; Ratner, M. A.; Wasielewski, M. R. J. Phys. Chem. A 2008,
112 (18), 4194–4201.
(80) Compton, R. H.; Grattan, K. T. V.; Morrow, T. J. Photochem. 1980,
14 (1), 61–66.
9
J. AM. CHEM. SOC. VOL. 131, NO. 48, 2009 17659

A R T I C L E S
Scott et al.
Table 1. Rate Constants for CS (kCS) and CR (kCR), the Spin-Spin Exchange Interaction (2J) Measured by MFE and TREPR, and the Rate
Constants for CR to the Ground State (kCRS) and to the Triplet State (kCRT) Determined from the MFE Data
a
)
b
)
d
)
d
compd
kCS (s-1
)
kCR (s-1
kCR (s-1
2JMFE (mT)
kCRS (s-1
kCRT (s-1
)
e
e
1
2
3
4
1.5 × 1011 ((0.3)
5.0 × 1010 ((0.5)
1.0 × 1010 ((0.05)
1.7 × 109 ((0.1)
1.3 × 108 ((0.4)
8.9 × 106 ((0.01)
2.9 × 106 ((0.04)
7.1 × 105 ((0.4)
1.3 × 108 ((0.4)
170 ( 1
30 ( 1
1.1 × 108
9.6 × 107
8.1 × 106
6.8 × 106 ((0.5)f
1.1 × 106 ((0.1)g
2 × 105
1.5 × 107 ((0.2)f
3.8 × 106 ((0.2)g
1.0 × 106
2.9 × 106
5.7 ( 1
7.5 × 105
0.3 ( 0.2c
0.9 ( 0.1
0.4 ( 0.1c
5
4.5 × 108 ((0.3)
2.7 × 105 ((0.3)
2.7 × 105
-
-
a CR rate constant measured by transient absorption kinetics monitoring NI-• decay at B ) 0 mT. b CR species associated rate constant. c Measured
by TREPR. d Determined from the kinetic analysis of the MFE data using khfc ) 1 × 108 s-1
.
. )
e For the fitting, krlx , 1 × 108 s-1 f For the fitting, krlx
1.1 × 106 ((0.5) s-1
.
g For the fitting, krlx ) 2.5 × 106 ((0.8) s-1
.
Figure 6. (A) Nanosecond transient absorption spectra of 3 in toluene at 295 K at the indicated times following a 7 ns, 416 nm laser pulse. Inset: transient
absorption kinetics of 3 in toluene at 295 K at 480 nm following a 7 ns, 416 nm laser pulse. Data points obtained every 5 nm. (B) Decay-associated spectra
and time constants of 3 obtained by global analysis of the transient absorption data.
Figure 7. Triplet and radical pair yields as a function of magnetic field (mT) for 1-4 in toluene at 480 nm at 295 K (130 ns delay for 3 and 860 ns delay
for 4). The insets show larger magnetic field ranges.
and 2, Figure 7. The 2J values for 1 and 2 decrease from 170
( 1 mT to 30 ( 1 mT, respectively, and show similar relative
triplet yields at 480 nm. The RP yield versus magnetic field
plots measured at 480 nm are displayed for 3 and 4 with 2J
resonances at 57 ( 1 mT and 3 ( 1 mT, respectively, Figure
7. The maximum of the relative triplet yield plot, or minimum
9
17660 J. AM. CHEM. SOC. VOL. 131, NO. 48, 2009

Spin-Selective Charge Transport Pathways
A R T I C L E S
Figure 8. Relative triplet yield, radical pair yield, and charge recombination
rate versus magnetic field for (A) 1 in toluene (75 ns delay after the laser
pulse for the RP) and (B) 2 in toluene (29 ns delay after the laser pulse for
the RP) at 480 nm.
Figure 9. TREPR spectra of (A) DMJ+•-An-Ph4-NI-• at 300 ns, (B) DMJ+•
-
An-Ph5-NI-• at 400 ns, following a 416 nm, 1.5 mJ laser pulse at the
indicated temperatures in toluene. Smooth curves superimposed on the
experimental spectra are computer simulations of the radical pair spectra
(see text) with the parameters given in Table 2. Positive features are in
enhanced absorption, while negative features are in emission.
of the relative RP yield plot correspond to the field at which
the S and T+1 (if 2J > 0) or T-1 (if 2J < 0) RP energy levels
cross. The magnitude of 2J is too small in compound 5 to be
accurately measured by MFE, so that TREPR experiments were
used to determine it. The magnetic field dependence of the triplet
yields of DMJ-3*An-Phn-NI measured at 430 nm are nearly
identical to those determined at 480 nm and are shown for 2 (n
) 2) in Figure S4.
Table 2. Exchange and Dipolar Interaction Parameters
compd
temp (K)
2J (mT)
D (mT)
4a
295
85
295
85
0.86 ( 0.05
0.80 ( 0.05
0.37 ( 0.05
0.28 ( 0.05
-
-0.18 ( 0.05
-
5b
-0.12 ( 0.05
Figure 8A,B shows plots of the relative RP and triplet yields
as well as the CR rates versus magnetic field for 1 and 2 in
toluene, respectively. These data qualitatively reveal for 1 that
the singlet CR pathway dominates and the total CR rate
decreases with increased triplet yield at the resonance condition.
Using the same analysis for 2, the opposite case is observed,
and the triplet CR pathway is more efficient than the singlet
CR pathway because both the total CR rate and the relative
triplet yield increase at the 2J resonance, while the RP yield is
minimized at resonance. The dominant recombination pathways
for 3 and 4 are similar to that of 2 and also show the rate and
triplet yield increases at resonance, while the RP yields decrease
(see Discussion).
a TREPR spectra obtained 300 ns after the laser pulse. b TREPR
spectra obtained 400 ns after the laser pulse.
TREPR. The RP spectra of 4 and 5 were simulated with the
SCRP mechanism using the model of Till and Hore.48 The best
fits to the data (Figure 9) used the parameters that are
summarized in Table 2. The calculated hyperfine coupling
constants of DMJ+•-An, as well as the measured g-factor and
hyperfine coupling constants of NI-• were used in the simula-
tions.66 The electron spin polarization pattern of the EPR signal,
i.e., which transitions are in enhanced absorption (A) or emission
(E), is determined by the SCRP sign rule,
Γ ) µ × sign[J-D(3 cos2(ꢂ) - 1)] )
Photoexcitation of 2-5 at 416 nm yields spin-polarized RPs
at both 295 and 85 K, indicating that charge transfer followed
by RP-ISC occurs at both temperatures. Figure 9 shows the
TREPR spectra of the RPs formed in 4 and 5 at 295 and 85 K
in toluene. The RP lifetime for 1 is too short to observe by
(-1) gives E/A or ) (+1) gives A/E (2)
where µ is -1 or +1 for a singlet or triplet excited state
precursor, respectively, ꢂ is the angle between the dipolar axis
of the radical pair and the direction of the magnetic field B0,
and D is the dipolar interaction between the two spins. Given
that photoexcitation initially produces a singlet RP, and that
the experimentally determined absolute phases of the RP spectra
are E/A,81,82 it follows from eq 2 that 2J is positive for the RPs
at all temperatures. The dipolar interaction (D) between the
(81) Adrian, F. J. J. Chem. Phys. 1972, 57 (12), 5107–5113.
(82) Atherton, N. M.; Davies, A. G. Chem. Soc. ReV. 1993, 22 (5), U293–
U293.
(83) Dance, Z. E. X.; Mi, Q.; McCamant, D. W.; Ahrens, M. J.; Ratner,
M. A.; Wasielewski, M. R. J. Phys. Chem. B 2006, 110 (50), 25163–
25173.
9
J. AM. CHEM. SOC. VOL. 131, NO. 48, 2009 17661

A R T I C L E S
Scott et al.
Figure 10. (A) Plots of kCS (9), kCR (2) at 295 K vs distance, rDA. The red lines are the linear fits to the data. (B) Logarithmic plot of the spin-spin
exchange interaction, 2J, versus distance rDA. (C) Plot of kCRT (9), kCRS (2) at 295 K vs distance, rDA. The error bars on the data points are smaller than the
size of the symbols for all plots.
electron spins in 2-5 is averaged out at room temperature by
molecular rotations and was estimated at 85 K from the
simulations (Table 2). The experimental details of the absolute
phase determination have been described previously.83
Our previous TREPR measurements66 on 1-4 at 85 K show
that CR within 3(DMJ+•-An-Phn-NI-•) results in the preferential
formation of DMJ-An-Phn-3*NI followed by exponentially
distance dependent TEnT in the range of 103-106 s-1 to yield
DMJ-3*An-Phn-NI. TEnT has been shown to occur by an
exchange mechanism that depends exponentially on the distance
between the donor and acceptor.84-86 The transient absorption
data at 295 K presented here show that (DMJ-An-Phn-3*NI) and
(DMJ-3*An-Phn-NI) are both derived from CR within 3(DMJ+•-
An-Phn-NI-•) with a total rate of kCRT; however, the TEnT rates
for 1-5 cannot be determined readily at 295 K due to
bimolecular annihilation in fluid solution.
of the superexchange mechanism in which the ꢀ values
depend on both the electronic coupling matrix element for
the charge transfer process, VDA, as well as the energy gap
for charge injection from the donor to the virtual bridge state,
24,87
∆EDB,
N-1
VDBVBA VBB
VDA
)
(4)
(
)
∆EDB ∆EDB
where ꢀ is described by
ꢀ ) ln
∆EDB
2
(5)
(
)
r
VBB
and where VDB and VBA are the matrix elements that couple
the donor to the bridge and the bridge to the acceptor,
respectively, VBB is the electronic coupling between bridge
sites, N is the number of identical bridge sites, and r is the
length of one bridge segment. Equation 4 is approximate and
does not take into account non-nearest neighbor interactions
and multiple pathways.88
Discussion
Superexchange-Mediated Charge Transport. The CS and CR
rate constants for 1-5 (Figure 10A) decay exponentially with
donor-acceptor distance as described by eq 3
Since the redox potentials for most conjugated bridge
molecules, including p-oligophenylenes, depend on bridge
length, the charge injection energy gap is also distance depend-
ent. As the bridge length increases, the energy gap may become
k ) k0e-ꢀ(r-r )
(3)
0
where k0 is the rate constant at the van der Waals contact
distance r0 (3.5 Å), and the measured value of ꢀ ) 0.35 Å-1
for CS and 0.34 Å-1 for CR. This behavior is characteristic
(86) Closs, G. L.; Johnson, M. D.; Miller, J. R.; Piotrowiak, P. J. Am.
Chem. Soc. 1989, 111 (10), 3751–3.
(84) Dexter, D. L. J. Chem. Phys. 1953, 21, 836–50.
(85) Closs, G. L.; Piotrowiak, P.; MacInnis, J. M.; Fleming, G. R. J. Am.
Chem. Soc. 1988, 110 (8), 2652–3.
(87) Ratner, M. A. J. Phys. Chem. 1990, 94 (12), 4877–4883.
(88) Onuchic, J. N.; Deandrade, P. C. P.; Beratan, D. N. J. Chem. Phys.
1991, 95 (2), 1131–1138.
9
17662 J. AM. CHEM. SOC. VOL. 131, NO. 48, 2009

Spin-Selective Charge Transport Pathways
A R T I C L E S
Table 3. Comparison of Ion Pair Distances and Values of 2J
rDA (Å) 2J (mT)
DMJ+•-An-Phn-NI-•
sufficiently small to allow oxidation or reduction of the bridge,
so that crossover from a superexchange to a charge hopping
mechanism may occur. The data presented here for DMJ-An-
Phn-NI do not show a switch in mechanism to the charge
hopping regime even at longer bridge distances, contrary to what
we observed earlier when p-oligophenylenes (n ) 1-5) are
linked to a phenothiazine (PTZ) donor and a perylene-3,4:9,10-
bis(dicarboximide) (PDI) electron acceptor. The PTZ-Phn-PDI
series shows a change in mechanism from superexchange to
n
PTZ+•-Phn-PDI-•
DMJ+•-An-Phn-NI-•
PTZ+•-Phn-PDI-•
1
2
3
4
5
16.5
20.9
25.5
29.9
34.3
12.8
17.1
21.4
25.7
30.0
170
31
5.7
0.9
0.4
-
170
31
6.4
1.5
hopping for CS at n ) 5 and CR at n ) 4.34 In the CS process,
+•
hole injection from 1*PDI into the Ph5 bridge forms PTZ-Ph5
-
tained for the DMJ-An-Phn-NI series are nearly identical to the
2J values measured for the PTZ-Phn-PDI series for the same
D-A distances and both decay exponentially with distance to
give R ) 0.36 Å-1 and 0.37 Å-1, respectively, which indicates
that the donor-acceptor coupling in these two p-oligophenylene
systems is similar. For the DMJ-An-Phn-NI series, the R value
PDI-•, while for the corresponding CR process, hole transfer
from PTZ+• to Ph4 or Ph5 forms the same type of intermediate.
This is a consequence of the increased ease of p-oligophenylene
bridge oxidation for longer bridges and the increased RP energy
of PTZ+•-Phn-PDI-• resulting from Coulomb destabilization as
n becomes larger. In contrast, CS within the DMJ-An-Phn-NI
series by the hopping mechanism requires reduction of the
bridge, DMJ+•-An-•-Phn-NI f DMJ+•-An-Phn-•-NI, which is
least endoergic for n ) 5 at ∆G ) 0.68 eV (Table S1). The
corresponding CR reaction may involve either bridge oxidation,
DMJ+•-An-Phn-NI f DMJ-An-Phn+•-NI-• or bridge reduction,
DMJ+•-An-Phn-NI-• f DMJ+•-An-Phn-•-NI. When n ) 5, the
free energies for bridge oxidation and reduction are ∆G ) 0.59
and 1.20 eV, respectively (Table S1). Thus, the charge hopping
mechanism for both CS and CR is precluded on thermodynamic
grounds, so that both reactions are expected to proceed
exclusively by the superexchange mechanism.
also closely matches the ꢀ value, which is reasonable given
93,94
that electron transfer theory shows that kCR ∝ VC2 R
,
while
eq 6 shows that 2J ∝ V2CRimplying that the prevalent charge
transport pathways in the DMJ-An-Phn-NI system involve only
superexchange.
Bridge Structure and Charge Recombination Dynamics. An
important aspect of analyzing the effect of the bridge properties
on the CR rate is to determine whether An serves as essentially
another phenyl group in the bridge or its extended π system
has a more significant influence on the bridge properties. Our
TREPR studies indicate that the radical cation spin density for
DMJ+•-An-Phn-NI-• is confined within DMJ with no delocal-
ization onto An. The ion pair energies calculated using eq 1
(Figure 5) are consistent with this observation because they show
that DMJ-An+•-Phn-NI-• is at least 0.33 eV above the fully
charge separated DMJ+•-An-Phn-NI-• state (n ) 1, Figure 5)
and is thus a virtual state with respect to the CR reaction.
Experimental evidence along with electronic structure calcula-
tions on conjugated bridges show that VDA for donors and
acceptors attached to the 9- and 10-positions of An is signifi-
cantly smaller than that for the corresponding Ph bridges largely
because steric hindrance provided by the peri-hydrogen atoms
adjacent to the 9- and 10-positions of An results in large
dihedral angles between the relevant π systems.95 An analysis
of the An-Ph dihedral angles compared to those of Ph-Ph
within the p-oligophenylene bridges reveals larger dihedral
angles for the An-Ph connection, Figure S5.
A comparison between the electronic coupling matrix ele-
ments for superexchange charge transfer through Phn and An-
Phn can be made by examining the measured 2J values for CR
in the DMJ+•-An-Phn-NI-• and PTZ+•-Phn-PDI-• series (Table
3). Note that the donor-acceptor distances for the molecules
having the An-Phn bridges are nearly the same as those for the
molecules having Phn+1 bridges. Interestingly, the 2J values for
CR within molecules having the An-Phn bridges are very similar
to those for the molecules having Phn+1 bridges. These data
strongly suggest that the superexchange interaction is propagated
through the An-Phn molecule in a manner similar to the Phn+1
molecule. The presence of the extended π system within An
does not increase the coupling, and the influence of the σ bonds
that directly link An and Ph should be considered. Recent studies
of charge transfer rates through a series of benzoannulated
The relationship between VDA and the spin-spin exchange
interaction, 2J, between the singlet and triplet RP states was
originally proposed by Kramers89 and developed further by
Anderson.41,90 The magnitude and sign of 2J depends on the
interactions of these states with other energetically nearby states
having the same respective spin multiplicities.41,90 For the CR
reaction, an approximate relationship between 2J and the
electronic coupling matrix elements for the singlet and triplet
pathways, VCRS and VCRT, respectively, is given by eq 643,54,91
V2CRT
V2CRS
2J )
-
(6)
∆GCRT + λ
∆GCRS + λ
where ∆GCRT and ∆GCRS are the free energies for charge
recombination to the triplet and singlet RPs, respectively, and
λ is the total reorganization energy for CR given by λ ) λi +
λs where λi is the internal reorganization energy for CR and λs
is the solvent reorganization energy. The value of λ is assumed
to be spin independent. The value of 2J depends exponentially
on the distance between two radical ions and is described by92
2J ) 2J0e-R(r-r )
(7)
0
where 2J0 is the spin-spin exchange interaction at the van der
Waals contact distance r0 of 3.5 Å and R is a constant. Figure
10B shows a plot of 2J vs distance that yields the distance decay
parameter, R ) 0.36 Å-1, which according to eq 6 directly
monitors the decay of the superexchange interaction through
the p-oligophenylene bridge. Interestingly, the 2J values ob-
(89) Kramers, H. A. Physica 1934, 1, 182–192.
(90) Anderson, P. W. Phys. ReV. 1950, 79 (2), 350–356.
(91) Nelsen, S. F.; Ismagilov, R. F.; Teki, Y. J. Am. Chem. Soc. 1998,
120, 2200–2201.
(93) Marcus, R. A. J. Chem. Phys. 1965, 43, 679–701.
(94) Jortner, J. J. Chem. Phys. 1976, 64 (12), 4860–4867.
(95) Lambert, C.; Risko, C.; Coropceanu, V.; Schelter, J.; Amthor, S.;
Gruhn, N. E.; Durivage, J. C.; Bredas, J. L. J. Am. Chem. Soc. 2005,
127 (23), 8508–8516.
(92) Kanter, F. J. J. D.; Kaptein, R.; Vansanten, R. A. Chem. Phys. Lett.
1977, 45 (3), 575–579.
9
J. AM. CHEM. SOC. VOL. 131, NO. 48, 2009 17663

A R T I C L E S
Scott et al.
S degenerate with T+1 or T-1 for positive or negative J values,
respectively. The polarization pattern of the SCRP spectra
observed by TREPR reveals that J is positive regardless of n,
which is not a factor in the kinetic analysis.66 At resonance,
S-T+1 mixing caused by the hyperfine interactions is considered
to be coherent within the characteristic time scale of the effective
hyperfine interaction, aeff ∼ 108 s-1. This fast mixing can be
described approximately as a kinetic equilibrium, if the other
processes are slow relative to the mixing time scale.97 At
resonance, we also need to consider S-T and T-T relaxation
processes.98 For simplicity, we assume that the relaxation rates,
krlx, are the same for S-T0 and for single quantum T0-T(1
relaxation as is the case at zero field. Double quantum relaxation
and S-T-1 relaxation can be ignored considering the large energy
gaps between the corresponding states. This approximation does
not cause serious problems because the main population flows
through fast S-T+1 mixing followed by recombination from both
the S and T+1 states unless relaxation is very fast.
At high fields (Figure 11C), we can ignore the effect of S-T(1
and single quantum T-T relaxations. The valid limits of this
treatment are confirmed by plotting the RP decay rate as a
function of the applied field. According to current ideas
regarding the relaxation mechanism,98,99 krlx is slower at higher
magnetic field, except when ultrahigh magnetic fields are
applied.100 The rate vs field plots for 1-4 (Figure 7) show flat
features at high field, where the field dependent relaxation rate
is much slower than the observed RP decay rate. Under these
conditions, the only spin states involved are S and T0, which
are mixed by incoherent relaxation.
Figure 11. Kinetic model of RP spin dynamics for CR in 2 and 3 for
(A) B ) 0, (B) B ) 2J, and (C) B . 2J. S, T+1, T0, and T-1 denote
corresponding RP spin sublevels. The dotted line represents the lowest
excited triplet state whose actual energy is far below that of the RP
energy levels. The circle indicates that the initial population is located
on the singlet RP.
bicyclo[2.2.2]octanes show that weakly coupled π pathways do
not increase electron transfer rates and that the dominant
pathway for charge transfer occurs through the σ system.22
Spin-Selective Charge Transport Pathways. The energy level
scheme depicted in Figure 1 shows that the CR within the singlet
and triplet RP states give different products. The rates at which
these products are formed and consequently their yields depend
on two major considerations: (1) Spin dynamics controls the
interconversion between the singlet and triplet RPs and deter-
mines population flow between the singlet and triplet manifolds.
(2) The CR rates within the singlet and triplet manifolds depend
on the electronic coupling matrix elements, VCRS and VCRT, as
well as the Franck-Condon weighted density of states, which
involves the reaction free energy and the total reorganization
energy for CR. The rates and yields of singlet and triplet CR
products in 1-5 are analyzed using these two basic ideas.
We have analyzed magnetic field dependent kinetic traces
for 1-4 based on the radical pair mechanism with spin state
dependent recombination reactions. For 1-3, we have observed
a distinct 2J resonance for both the triplet yield and the RP
decay rate. The spin dynamics of the RP, where 2J is larger
than the effective hyperfine interaction (aeff ∼ 3mT),66,96 are
described in Figure 11. For all cases, the initial RP population
is localized only on the singlet state, and the kinetic equations
used to obtain the fits are outlined in the Supporting Information.
The CS kinetics are not considered because they are much faster
than the CR kinetics and are not influenced by spin dynamics.
In the absence of a magnetic field (Figure 11A), the three
triplet sublevels of the RP are nearly degenerate and are split
from the singlet level by 2J. Since the CR reactions are spin
selective, the singlet RP exclusively recombines to the singlet
ground state and the triplet RP recombines to the neutral triplet
excited state with rate constants of kCRS and kCRT, respectively.
Note that kCRT is understood here as the sum of the recombina-
tion rates to 3*NI and 3*An. The detailed RP-ISC process at low
fields for systems with large 2J has never been explicitly treated,
but it is likely that the mixing process is governed by incoherent
relaxation rather than coherent interconversion of the singlet
and triplet RP states, considering the large energy gap (2J) and
small hyperfine coupling.
For 1, the kinetics at zero field and at high field are almost
the same, both of which show negligible triplet yield. This
feature indicates that S-T0 relaxation is much slower than the
RP lifetime because 2J is very large and the radical pair lifetime
is short. In this case, the mechanistic scheme simplifies; at
magnetic fields far from resonance, the apparent decay rate of
the RP should be identical to kCRS. At resonance, the decay rate
becomes kobs ) kCRS/2 + kCRT/2 because of fast interconversion
only between S and T+1. The monoexponential fits to the data
for 1 at zero and high fields using the analytical solution for
kobs give almost identical values of kCRS and kCRT as described
in Table 2 and Figure 10C. It is difficult to solve the equations
describing the kinetics for 2 and 3 analytically, so we have
solved them numerically using a matrix formalism.97 We pick
three kinetic traces: at zero field, at resonance, and at high field
and fit them by simulation (Supporting Information, Figure S6).
The parameters obtained by this method are summarized in
Table 1. The rate constants kCRS and kCRT are obtained with a
small error for both 2 and 3. The relaxation rate constant krlx
for 2 is slower than that for 3 (footnotes to Table 1), which is
reasonable given that the energy gap is larger for 2 than for 3.
However, the detailed mechanism of S-T0 relaxation is still an
open question. One possibility could be the interplay of the
hyperfine interaction and the very fast singlet-triplet dephasing
(97) Murakami, M.; Maeda, K.; Arai, T. J. Phys. Chem. A 2005, 109
(26), 5793–5800.
(98) Hayashi, H.; Nagakura, S. Bull. Chem. Soc. Jpn. 1984, 57 (2), 322–
328.
At resonance (Figure 11B), where B0 ) 2J, the Zeeman
splitting of triplet sublevels exactly matches 2J, which makes
(99) Carrington, A.; McLachlan, A. D. Introduction to magnetic resonance
with applications to chemistry and chemical physics; Harper & Row:
New York, 1967.
(96) Hayashi, H. Introduction to Dynamic Spin Chemistry: Magnetic Field
Effects upon Chemical and Biochemical Reactions; World Scientific:
Singapore, 2004.
(100) Fujiwara, Y.; Aoki, T.; Yoda, K.; Cao, H.; Mukai, M.; Haino, T.;
Fukazawa, Y.; Tanimoto, Y.; Yonemura, H.; Matsuo, T.; Okazaki,
M. Chem. Phys. Lett. 1996, 259 (3-4), 361–367.
9
17664 J. AM. CHEM. SOC. VOL. 131, NO. 48, 2009

Spin-Selective Charge Transport Pathways
A R T I C L E S
1
3
Figure 12. Left: CR pathways for (DMJ+•An-Phn-NI-•). Right: CR pathways for (DMJ+•An-Phn-NI-•).
caused by fluctuations of the exchange interaction, 2J.101,102
These fluctuations could be a result of torsional motions about
the single bonds joining each phenyl, as well as joining the
bridge to the donor and acceptor.
rate vs free energy curve103 because |∆GCRT| < 0.40 eV and is
faster than singlet CR 1(DMJ+•-An-Phn-NI-•) f DMJ-An-Phn-
NI, which occurs deep within the Marcus inverted region (∆GCRS
) -2.1 to -2.4 eV for n ) 1-5).9
For 4, which shows a negligible resonance in its MFE plot
due to the fact that 2J , aeff, the spin dynamics can be explained
in terms of conventional hyperfine and relaxation mechanisms.49,96
At zero field, all four RP states are completely mixed by the
hyperfine interaction and the total rate is
The separation of the rate constants for the singlet and triplet
recombination pathways allows us to further determine the
dominant charge transport mechanism as given by ꢀ, Figure
10C. The rate vs distance plots for the singlet and triplet
pathways show that a superexchange mechanism prevails for
CR with ꢀ ) 0.35 Å-1 for kCRT and ꢀ ) 0.48 Å-1 for kCRS
.
1
4
3
4
Since the triplet CR rates are similar to the total CR rates, it is
reasonable that ꢀ ) 0.34 Å-1 for the total CR rate as measured
by the decay of NI•- matches ꢀ ) 0.35 Å-1 for the triplet
pathway. This is also in agreement with the results obtained
from the distance dependence of 2J, where R ) 0.36 Å-1, since
eq 6 predicts 2J values that are dominated by the triplet pathway
(∆GCRT + λ , ∆GCRS + λ). The larger ꢀ value observed for
the singlet CR pathway relative to that for the triplet pathway
can be explained by considering the various electron transfer
pathways available. For CR within 1(DMJ+•An-Phn-NI-•),
moving an electron via the HOMOs (Figure 12, left) must
produce 1*NI (blue arrows), whereas moving an electron via
the LUMOs must produce either 1*DMJ or the ground state of
the system (red arrow). We do not observe 1*NI, but we do
observe CR to the ground state, so that singlet CR only involves
kobs(0mT) ) kCRS
+
kCRT
(8)
At high field, where the rate vs field plot is flat, only the S and
T0 states are fully mixed and the effect of relaxation between
S-T0 mixed state and T(1 state can be ignored. Thus, the rate
becomes
1
2
1
2
kobs(high field) ) kCRS
+
kCRT
(9)
The analytical solution of the two equations using the observed
rates at zero and high fields gives kCRS and kCRT as shown in
Table 1 and Figure 10C.
The similarity of the overall CR rates (kCR) at zero field
(Figure 11A) to those of kCRT is explained by the spin dynamics
of the RP, which is largely dependent on the 2J energy gap.
When 2J is much larger than aeff (n ) 1, 2), CR occurs mainly
by the singlet pathway because the singlet and triplet RP pair
states interconvert slowly (see krlx for 2). On the other hand,
triplet CR predominates for 3 and 4 because of their smaller
S-T energy gaps. Since S-T interconversion is faster than kCRS
(see krlx for 3) and kCRT is also fast, the value of overall rate
constant kCR is close to that of kCRT rather than kCRS. RP-ISC is
governed by subtle spin interactions (,1 cm-1) and thus can
be the rate-determining step in the overall CR process when 2J
is sufficiently large and the condition kCRS , krlx , kCRT is
fulfilled.
3
the LUMOs. In contrast, considering CR for (DMJ+•An-Phn-
NI-•), moving an electron via the HOMOs (Figure 12, right)
must produce 3*NI (blue arrows), whereas moving an electron
via the LUMOs must produce 3*DMJ, which would undergo
triplet-triplet energy transfer very rapidly to produce 3*An (red
arrow). We observe both 3*NI and 3*An following CR at 295
K, so that CR of the triplet RP to form 3*NI involves the
HOMOs, while CR to form 3*An involves the LUMOs. Since
superexchange involving the LUMOs has larger energy gaps,
∆EDB, than does superexchange involving the HOMOs, eq 5
predicts that ꢀ will be larger for singlet CR than triplet CR, as
we observe.
Temperature Dependence of Charge Recombination. As
mentioned above, CR at 295 K initially produces both 3*NI and
3*An, while CR at 85 K produces 3*NI preferentially. On the
basis of the orbital discussion in the previous section, the room
temperature result suggests that charge transfer uses both the
HOMO and LUMO orbitals of the bridge, while the low
temperature result suggests that charge transfer uses the HOMO
While spin dynamics governs the populations of singlet and
triplet RPs, the subsequent CR reactions within these manifolds
depend on the usual considerations of electron transfer theory.93,94
The triplet CR 3(DMJ+•-An-Phn-NI-•) f DMJ-An-Phn-3*NI and
DMJ-3*An-Phn-NI occur within the normal region of the Marcus
(101) Miura, T.; Maeda, K.; Arai, T. J. Phys. Chem. A 2006, 110 (12),
4151–4156.
(102) Miura, T.; Murai, H. J. Phys. Chem. A 2008, 112 (12), 2526–2532.
(103) Marcus, R. A. J. Chem. Phys. 1956, 24, 966–78.
9
J. AM. CHEM. SOC. VOL. 131, NO. 48, 2009 17665

A R T I C L E S
Scott et al.
orbitals leading to preferential production of 3*NI. An examina-
tion of the data in Figure 5 shows that the DMJ-An-Phn+•-NI-•
virtual bridge states involving the bridge HOMOs have energies
that are closer to those of DMJ+•-An-Phn-NI-• than are the
energies of the corresponding states DMJ+•-An-Phn-•-NI involv-
ing the bridge LUMOs. Thus, based on the energy gap
dependence (∆EDB) given in eq 4, the electronic coupling for
triplet CR (VCRT) leading to 3*NI should be somewhat larger
than that leading to 3*An. Using the functional form of the
semiclassical expression for the electron transfer rate (eq
10),93,94,104
An-Ph2-3*An, ∆GCRT ) -0.40 eV at 295 K and -0.60 eV at
85 K. Using these values of ∆GCRT, λi ) 0.39 eV, and λS )
0.07 eV; and assuming that VCRT is the same at both 295 and
NI
NI
85 K, eq 10 yields k /kAn ) 1.0 at 295 K and k /kAn
)
CRT CRT
CRT CRT
1.7 at 85 K. Thus, for 2 at room temperature, CR recombination
to 3*NI and 3*An proceed at similar rates, while at 85 K CR to
3*NI is nearly twice as fast as CR to 3*An. This agrees with our
previous TREPR results at 85 K, which show a preference of
hole transfer to initially form 3*NI, rather than electron transfer
to form 3*An, as well as our room temperature results presented
here, which do not show a strong preference.
Conclusions
kCRT
)
(∆GCRT + λs + npω)2
4λskBT
∞
We have prepared a series of D-B-A molecules to understand
the charge transport properties in simple bridging structures
consisting of p-oligophenylenes (n ) 1-5) covalently attached
to a DMJ-An donor and an NI acceptor. We have determined
how the electronic coupling depends on molecular structure by
analyzing the spin-spin exchange interaction (2J) between the
unpaired spins of the radical pair as determined by magnetic
field effects on triplet and radical ion pair yields as well as time-
resolved EPR spectroscopy and have established that the
distance dependence of charge recombination in D-B-A mol-
ecules is spin-selective using magnetic field effect data and a
kinetic model to separate the spin-selective charge recombination
rates, kCRS and kCRT. The spin dynamics between the radical
ion pairs and the electron transfer parameters within each spin
manifold combine to strongly favor triplet charge recombination
that results in the formation of both 3*An and 3*NI. An analysis
of these phenomena in the context of superexchange theory leads
us to identify participation of the bridge HOMOs in the
formation of 3*NI, while participation of the bridge LUMOs
leads to formation of 3*An. A similar analysis also explains why
charge recombination leads preferentially to 3*NI formation at
low temperatures and comparable amounts of 3*NI and 3*An at
room temperature. Interestingly, the structural and electronic
impact of An integrated within the p-oligophenylene bridge does
not significantly change the donor-acceptor electronic coupling
within this system. These results show that once CS has
occurred, An behaves like Ph with the bridge. It is possible
that the dihedral angles between An and Phn are sufficiently
large that the propagation of electronic coupling through the
An-Phn bridge involves both the σ and π systems of An.
Developing systems for solar energy conversion and organic
electronics requires a thorough knowledge of the dependence
of charge transport properties on molecular structure. In addition,
the results presented here show that the spin dynamics of the
system provide an important tool for understanding charge
transport mechanisms as well as a potential means of controlling
long-distance charge transport in organic materials.
n
2π
p
1
2
-
e(-S)S /n!
e
(10)
[
]
|VCRT
|
∑
4πλ k T
ꢀ
s
B
n)0
where pω is the vibrational quantum, assumed to be 1500 cm-1
and S ) λi/pω, a larger value of VCRT should result in a faster
triplet CR rate. However, lowering the temperature to 85 K, at
which toluene is a frozen glass, usually results in changes to
∆GCRT and λS as well as VCRT. We have demonstrated that
restricting solvent dipole rotations by freezing the solvent results
in large increases in ion pair energies.105 These energies increase
by as much as 0.75 eV relative to those obtained at room
temperature in highly polar fluid media and about 0.2 eV above
those obtained in liquid toluene. Moreover, since the frozen
solvent behaves essentially as a low polarity medium, λS = 0
as determined from the Marcus treatment based on dielectric
continuum theory.103 Thus, given that λS is very small for CR
in 1-5 in toluene at 295 K, λS is nearly temperature indepen-
dent. Assuming that the ion pair energies of DMJ+•-An-Phn-
NI-•, DMJ-An-Phn+•-NI-• and DMJ+•-An-Phn-•-NI (Table S1)
all increase by about 0.2 eV in glassy toluene at 85 K relative
to their values determined in fluid toluene at 295 K, the
corresponding values of ∆EDB for the HOMO and LUMO
superexchange pathways leading to 3*NI and 3*An, respectively,
do not change, so that based on eq 4 VCRT does not change, and
the HOMO pathway is still favored. This analysis is supported
by the fact that the TREPR measurements on DMJ+•-An-Phn-
NI-• in 4 and 5 show that 2J decreases only slightly as the
temperature is lowered from 293 to 85 K (Figure 9, Table 1).
Thus, the values of VCRT for 4 and 5 only decrease by a small
amount as the temperature decreases, most likely due to a
reduction in the amplitude of torsional motions about the single
bonds within the p-oligophenylene bridges.83,106
Despite the weak temperature dependence of VCRT, an
additional change in the relative rates of triplet CR leading to
3*NI and 3*An may result from the temperature dependence of
the Franck-Condon weighted density of states (FCWD) term
in eq 10. As noted above, the values of ∆GCRT for the CR
reactions that produce 3*NI and 3*An from (DMJ+•-An-Phn-
3
Acknowledgment. This work was supported by the Chemical
Sciences, Geosciences, and Biosciences Division, Office of Basic
Energy Sciences, DOE, under grant no. DE-FG02-99ER14999. The
authors also thank Prof. Emily Weiss and Dr. Raanan Carmieli for
many helpful discussions.
NI-•) become more negative as the temperature is lowered,
largely as a consequence of destabilization of the ion pair state
upon freezing the solvent.105 For example, for (DMJ+•-An-
3
Ph2-NI-•) f DMJ-An-Ph2-3*NI, ∆GCRT ) -0.22 eV at 295 K
and -0.42 at 85 K, while for (DMJ+•-An-Ph2-NI-•) f DMJ-
3
Supporting Information Available: Details regarding the
synthesis and characterization of 5, and additional redox,
transient absorption, and magnetic field effect data. This material
is available free of charge via the Internet at http://pubs.acs.org.
(104) Hopfield, J. J. Proc. Natl. Acad. Sci. U.S.A. 1974, 71, 3640–3644.
(105) Gaines, G. L., III; O’Neil, M. P.; Svec, W. A.; Niemczyk, M. P.;
Wasielewski, M. R. J. Am. Chem. Soc. 1991, 113 (2), 719–21.
(106) Berlin, Y. A.; Grozema, F. C.; Siebbeles, L. D. A.; Ratner, M. A. J.
Phys. Chem. C 2008, 112 (29), 10988–11000.
JA907625K
9
17666 J. AM. CHEM. SOC. VOL. 131, NO. 48, 2009