Unexpectedly similar charge transfer rates through benzo-annulated bicyclo[2.2.2]octanes

Article abstract of DOI:10.1021/ja8004623

A 4-(pyrrolidin-1-yl)phenyl electron donor and 10-cyanoanthracen-9-yl electron acceptor are attached via alkyne linkages to the bridgehead carbon atoms of bicyclo[2.2.2]octane and all three benzo-annulated bicyclo[2.2.2] octanes. The σ-system of bicyclo[2.2.2]octane provides a scaffold having nearly constant bridge geometry on which to append multiple, weakly interacting benzo π-bridges, so that the effect of incrementally increasing numbers of π-bridges on electron transfer rates can be studied. Surprisingly, photoinduced charge transfer rates measured by transient absorption spectroscopy in toluene show no benefit from increasing the number of bridge π-systems, suggesting dominant transport through the σ-system. Even more surprisingly, the significant changes in hybridization undergone by the σ-system as a result of benzo-annulation also appear to have no effect on the charge transfer rates. Natural Bond Orbital analysis is applied to both σ- and π-communication pathways. The transient absorption spectra obtained in 2-methyltetrahydrofuran (MTHF) show small differences between the benzo-annulated molecules that are attributed to changes in solvation. All charge transfer rates increase significantly upon cooling the MTHF solutions to their glassy state. This behavior is rationalized using combined molecular dynamics/electronic structure trajectories.

Full text of DOI:10.1021/ja8004623

Published on Web 05/24/2008
Unexpectedly Similar Charge Transfer Rates through
Benzo-Annulated Bicyclo[2.2.2]octanes
Randall H. Goldsmith,† Josh Vura-Weis,† Amy M. Scott,† Sachin Borkar,‡
Ayusman Sen,‡ Mark A. Ratner,*,† and Michael R. Wasielewski*,†
Department of Chemistry, Argonne-Northwestern Solar Energy Research (ANSER) Center, and
International Institute for Nanotechnology, Northwestern UniVersity, EVanston, Illinois 60208
and the Department of Chemistry, The PennsylVania State UniVersity, UniVersity Park,
PennsylVania 16802
Received January 20, 2008; E-mail: m-wasielewski@northwestern.edu; ratner@northwestern.edu
Abstract: A 4-(pyrrolidin-1-yl)phenyl electron donor and 10-cyanoanthracen-9-yl electron acceptor are
attached via alkyne linkages to the bridgehead carbon atoms of bicyclo[2.2.2]octane and all three benzo-
annulated bicyclo[2.2.2]octanes. The σ-system of bicyclo[2.2.2]octane provides a scaffold having nearly
constant bridge geometry on which to append multiple, weakly interacting benzo π-bridges, so that the
effect of incrementally increasing numbers of π-bridges on electron transfer rates can be studied. Surprisingly,
photoinduced charge transfer rates measured by transient absorption spectroscopy in toluene show no
benefit from increasing the number of bridge π-systems, suggesting dominant transport through the σ-system.
Even more surprisingly, the significant changes in hybridization undergone by the σ-system as a result of
benzo-annulation also appear to have no effect on the charge transfer rates. Natural Bond Orbital analysis
is applied to both σ- and π-communication pathways. The transient absorption spectra obtained in
2-methyltetrahydrofuran (MTHF) show small differences between the benzo-annulated molecules that are
attributed to changes in solvation. All charge transfer rates increase significantly upon cooling the MTHF
solutions to their glassy state. This behavior is rationalized using combined molecular dynamics/electronic
structure trajectories.
transfer6,7 and to possess a rich electronic structure,8–11 including
interference effects.12 More recently, the majority of attention
Introduction
Covalently linked Donor-Bridge-Acceptor (D-B-A) mol-
ecules, where the bridge is comprised of a σ-bond network, have
provided a means of elucidating many of the important aspects
of intramolecular charge and energy transfer. Saturated spacer
groups provided the first unambiguous confirmation of the
existence of the Marcus inverted region.1,2 Bicyclo[2.2.2]octane3
and norbornane4 spacers provided early means to determine the
distance dependence of intramolecular charge transfer through
the incremental increase in the length of an oligomeric bridge
group. The utility of these classes of molecules stems from their
role as nearly rigid spacers or molecular rods.5 Despite earlier
assumptions about their designation as inert spacers,6 systems
containing saturated hydrocarbon bridges have been clearly
shown to be electronically relevant to intramolecular charge
has moved away from looking at bridges having all σ-bonds to
those with π-bonds. Such a shift was motivated by the increased
ability of these π-bridges to convey electronic communication
between the π-systems of the D and A groups by maintaining
closer energetic resonance, suggested by McConnell’s expres-
sion for the effective electronic coupling (Veff) between D and
A,13 eq 1,
N-1
VD1VNA
Vi,i+1
Veff
)
(1)
∏
E1 - ED⁄A i)1 Ei+1 - ED⁄A
where Vij are matrix elements between the different subunit wave
functions, Ei is the energy of a particular subunit, and the bridge
has N equivalent subunits.
The pursuit of greater electronic communication has been
driven by the potential development of molecular electronic
† Northwestern University.
‡ The Pennsylvania State University.
(1) Miller, J. R.; Calcaterra, L. T.; Closs, G. L. J. Am. Chem. Soc. 1984,
106 (10), 3047.
(2) Wasielewski, M. R.; Niemczyk, M. P.; Svec, W. A.; Pewitt, E. B.
J. Am. Chem. Soc. 1985, 107 (19), 5562.
(7) Zimmerman, H. E.; McKelvey, R. D. J. Am. Chem. Soc. 1971, 93
(15), 3638.
(3) Leland, B. A.; Joran, A. D.; Felker, P. M.; Hopfield, J. J.; Zewail,
A. H.; Dervan, P. B. J. Phys. Chem. 1985, 89 (26), 5571.
(4) Oevering, H.; Paddon-Row, M. N.; Heppener, M.; Oliver, A. M.;
Cotsaris, E.; Verhoeven, J. W.; Hush, N. S. J. Am. Chem. Soc. 1987,
109 (11), 3258.
(8) Beratan, D. N. J. Am. Chem. Soc. 1986, 108 (15), 4321.
(9) Braga, M.; Larsson, S. Chem. Phys. Lett. 1993, 213 (3-4), 217.
(10) Liang, C. X.; Newton, M. D. J. Phys. Chem. 1992, 96 (7), 2855.
(11) Naleway, C. A.; Curtiss, L. A.; Miller, J. R. J. Phys. Chem. 1991, 95
(22), 8434.
(5) Schwab, P. F. H.; Levin, M. D.; Michl, J. Chem. ReV. 1999, 99 (7),
1863.
(12) Paddon-Row, M. N.; Shephard, M. J. J. Am. Chem. Soc. 1997, 119
(23), 5355.
(6) Paddon-Row, M. N. AdV. Phys. Org. Chem. 2003, 38, 1.
(13) McConnell, H. M. J. Chem. Phys. 1961, 35, 508.
9
10.1021/ja8004623 CCC: $40.75
2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 7659–7669 7659

A R T I C L E S
Goldsmith et al.
Figure 1. Structures of Donor-Bridge-Acceptor molecules with benzo-annulated BCO bridges.
devices (molecular wires)14 as well as understanding ultrafast
energy and electron transfer in the photosynthetic reaction center
and related systems.15 Studies of π-bridges have revealed
qualitatively new transfer regimes that depend on energetic
resonance conditions.16–20 Albinsson and co-workers have
recently shown a stark contrast between the efficacy of σ- and
π-type bridges by comparing a phenyl group and a bicyclo-
[2.2.2]octane group in equivalent positions in a D-B-A
system.21,22 Following the predictions of many earlier experi-
mental and theoretical studies, charge transfer was strongly
inhibited by the σ-bridge.
Other recent studies attempting to elucidate ultrafast charge
transfer in biological systems have suggested that multiple
spatial pathways,23–26 possibly summing coherently, may con-
tribute to the overall charge transfer rate. As part of our ongoing
efforts to understand charge transfer through multiple spatial
pathways,27–29 we have attempted to use the dichotomy between
σ- and π-type bridges to develop D-B-A molecules to test
predictions about the role of multiple spatial pathways in charge
transfer. The σ-system provides a scaffold having nearly constant
bridge geometry on which to append multiple weakly interacting
π-bridges, so that the effect of incrementally increasing numbers
of π-bridges on electron transfer rates can be studied.
For our σ-scaffold, we chose a 1,4-disubstituted bicyclo-
[2.2.2]octane (BCO) group. A 4-(pyrrolidin-1-yl)phenyl (PP)
electron donor and 10-cyanoanthracen-9-yl electron acceptor
(CA) were then attached via alkyne linkages to the scaffold.
π-Pathways were then added by replacing the BCO with benzo-
annulated BCOs. Up to three π-pathways could be added, with
the three-phenyl system comprised of a bridgehead substituted
triptycene, as seen in Figure 1. While a few examples of charge
and energy transfer systems using asymmetric bridgehead
substituted BCO3,7,30–32 and triptycene33 bridges have been
reported, and many reports have used various benzo-annulated
BCO molecules to investigate the relative importance of
through-space and through-bond effects,34,35 to our knowledge,
this is the first report to examine charge separation and
recombination across the full series of four doubly bridgehead
substituted benzo-annulated BCOs.
Experimental Section
(14) Cuniberti, G.; Fagas, G.; Richter, K. Introducing Molecular Electron-
ics; Springer Verlag: Berlin, 2005; Vol. 680.
Full details regarding the synthesis and characterization of all
new compounds can be found in the Supporting Information.
Steady-state absorption measurements were performed with a
Shimadzu (UV-1601) spectrophotometer. Femtosecond transient
absorption measurements were performed using the frequency
doubled output of a regeneratively amplified Ti:sapphire laser
operating at 2 kHz. The samples were irradiated with 1.0-1.2 µJ,
416 nm, 120 fs laser pulses focused to a 200 µm diameter spot.
The total instrument response time for the pump-probe experiment
was 180 fs. The absorbance of all room temperature samples for
femtosecond transient absorption spectroscopy was maintained
between 0.5 and 0.7 at 416 nm in a 2 mm path length cuvette.
Variable temperature studies were conducted using a Janis VNF-
100 cryostat with a Cryo-con 32B temperature controller and a
(15) Wasielewski, M. R. J. Org. Chem. 2006, 71 (14), 5051.
(16) Bixon, M.; Jortner, J. J. Chem. Phys. 1997, 107 (13), 5154.
(17) Davis, W. B.; Svec, W. A.; Ratner, M. A.; Wasielewski, M. R. Nature
1998, 396 (6706), 60.
(18) Davis, W. B.; Wasielewski, M. R.; Ratner, M. A.; Mujica, V.; Nitzan,
A. J. Phys. Chem. A 1997, 101 (35), 6158.
(19) Goldsmith, R. H.; Sinks, L. E.; Kelley, R. F.; Betzen, L. J.; Liu, W. H.;
Weiss, E. A.; Ratner, M. A.; Wasielewski, M. R. Proc. Natl. Acad.
Sci. U.S.A. 2005, 102 (10), 3540.
(20) 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.
(21) Kilså, K.; Kajanus, J.; Macpherson, A. N.; Mårtensson, J.; Albinsson,
B. J. Am. Chem. Soc. 2001, 123 (13), 3069.
(22) Pettersson, K.; Wiberg, J.; Ljungdahl, T.; Mårtensson, J.; Albinsson,
B. J. Phys. Chem. A 2006, 110 (1), 319.
(23) Balabin, I. A.; Onuchic, J. N. Science 2000, 290 (5489), 114.
(24) Kawatsu, T.; Kakitani, T.; Yamato, T. J. Phys. Chem. B 2002, 106
(43), 11356.
(30) Voegtle, F.; Frank, M.; Nieger, M.; Belser, P.; von Zelewsky, A.;
Balzani, V.; Barigelletti, F.; De Cola, L.; Flamigni, L. Angew. Chem.,
Int. Ed. Engl. 1993, 32 (11), 1643.
(25) Kuki, A.; Wolynes, P. G. Science 1987, 236 (4809), 1647.
(26) Prytkova, T. R.; Kurnikov, I. V.; Beratan, D. N. Science 2007, 315
(5812), 622.
(31) De Cola, L.; Balzani, V.; Barigelletti, F.; Flamigni, L.; Belser, P.;
Vonzelewsky, A.; Frank, M.; Vogtle, F. Inorg. Chem. 1993, 32 (23),
5228.
(27) Goldsmith, R. H.; Wasielewski, M. R.; Ratner, M. A. J. Am. Chem.
Soc. 2007, 129 (43), 3066.
(32) Zimmerman, H. E.; Goldman, T. D.; Hirzel, T. K.; Schmidt, S. P. J.
Org. Chem. 1980, 45 (20), 3933.
(28) Goldsmith, R. H.; Wasielewski, M. R.; Ratner, M. A. J. Phys. Chem.
B 2006, 110 (41), 20258.
(33) Beyeler, A.; Belser, P. Coord. Chem. ReV. 2002, 230 (1-2), 29.
(34) Adcock, W.; Trout, N. A. Chem. ReV. 1999, 99 (5), 1415.
(35) Stock, L. M. J. Chem. Educ. 1972, 49 (6), 400.
(29) Weiss, E. A.; Sinks, L. E.; Lukas, A. S.; Chernick, E. T.; Ratner,
M. A.; Wasielewski, M. R. J. Phys. Chem. B 2004, 108, 10309.
9
7660 J. AM. CHEM. SOC. VOL. 130, NO. 24, 2008

Similar Charge Transfer Rates through Benzo-annulation
A R T I C L E S
Scheme 1a
a Reagents: (a) KOC2H5, EtOH, 55%; (b) (1) NaH, DME, (2) C2H4Br2, 71%; (c) HS(CH2)3SH, BF3 ·Et2O, CHCl3, 80%; (d) Raney Ni, EtOH, 98%; (e)
LiAlH4, Et2O, 89%; (f) oxalyl chloride, DMSO, DCM, quant.; (g) CBr4, PPh3, Zn, DCM, hexanes, 53%; (h) n-BuLi, THF, 76%; (i) (1) N-(4-
iodophenyl)pyrrolidine, piperidine, CuI, Pd(PPh3)4, (2) bromocyanoanthracene, TEA, CuI, Pd(PPh3)2Cl2, 9%.
Scheme 2a
a Reagents: (a) TIPS acetylene, DEA, CuI, Pd(PPh3)2Cl2, 95%; (b) isoamyl nitrite, anthranilic acid, dioxane, 43%; (c) TBAF, THF, 73%; (d) (1)
bromocyanoanthracene, TEA, CuI, Pd(PPh3)2Cl2, (2) N-(4-iodophenyl)pyrrolidine, piperidine, CuI, Pd(PPh3)4, 15%.
procedures employed by Stock,42,43 Sukenik,44 Kurreck,45
home-built quartz sample holder with a 1.6 mm path length, so
Albinsson,46 and co-workers, Scheme 1. The dialkyne derivative
that the sample absorbance at 416 nm was maintained between 0.5
and 0.7. Nanosecond transient absorption measurements were made
using a Continuum Panther OPO pumped by the frequency-tripled
output of a Continuum 8000 Nd:YAG laser. The probe light in the
nanosecond experiment was generated using a xenon flashlamp
(EG&G Electrooptics FX-200) and detected using a photomultiplier
tube with high voltage applied to only four dynodes (Hamamatsu
R928). The total IRF is 7 ns and is determined primarily by the
laser pulse duration. The samples for the nanosecond experiments
had an absorbance of 0.5-0.7 at 416 nm in a 1 cm path length
cuvette. These samples were subjected to four freeze-pump-thaw
cycles and excited at either 416 or 430 nm using 2 mJ per pulse
(10 Hz) focused to a 5 mm diameter spot size. A full description
of the femtosecond transient absorption36 and nanosecond transient
absorption20 apparatus can be found elsewhere. Transient absorption
kinetics were fit to a sum of exponentials with a Gaussian instrument
response function by using Levenberg-Marquardt least-squares
fitting. Geometry optimizations and single-point calculations were
performed using the B3LYP functional with the 6-31G* basis set
with Gaussian 98.37 NBO analysis was performed using the NBO
3.1 add-on38 with the B3LYP functional and the 6-31G* basis
set. Molecular dynamics were performed with MOPAC 2007,39
employing electronic structure calculations at the AM1/CI level
with a 3 electron 2 orbital (HOMO, HOMO-1) active space.
of triptycene was prepared using a modification of the procedure
of Michl and co-workers,47 Scheme 2.
In contrast, the benzobicyclo[2.2.2]octene (BBCO) and
dibenzo-bicyclo[2.2.2]octadiene (DBCO) derivatives with alkynes
at the bridgeheads were unknown. The preparation of BBCO,
Scheme 3, started from dimethyl-1,3-cyclohexadiene-1,4 dicar-
boxylate, which can be produced in multigram quantities from
the corresponding saturated cyclohexyl derivative.48 The esters
were reduced following Chu’s procedure,49 and the alcohols
were immediately protected with a tert-butyldimethylsilyl
(TBDMS) group. Reduction of the esters was necessary at this
stage because they deactivated the ring system to a degree that
made the diene inert toward a Diels-Alder reaction with
benzoquinone and benzyne. The protected diol, however, reacted
easily with benzoquinone to form the BBCO core while an
attempted reaction with benzyne yielded a complex mixture of
(36) 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.
(37) Frisch, M. J. et al. Gaussian 98, revision A.7; Gaussian, Inc.:
Pittsburgh, PA, 1998. Full reference in Supporting Information.
(38) Glendening, E. D.; Reed, A. E.; Carpenter, J. E.; Weinhold, F. QCPE
Bull. 1990, 10, 58.
(39) Stewart, J. J. P. MOPAC 2007; (http://mopac2007.net).
(40) Effenberger, F.; Agster, W.; Fischer, P.; Jogun, K. H.; Stezowski, J. J.;
Daltrozzo, E.; Kollmannsberger-von Nell, G. J. Org. Chem. 1983, 48
(24), 4649.
Results
Molecular Design. The donor and acceptor in compounds 1-4
were chosen because of their excellent spectroscopic and redox
properties, as will be described later, and because they can
connect with the bridge via an alkyne linkage, which may
be capable of conveying electronic communication between the
D/A and the bridge π-system. The syntheses of 1-4 were
executed using a building-block approach. Aryl halides of the
donor40 and acceptor41 were prepared in one step from com-
mercial materials. The synthesis of the ethynylated BCO and
benzo-BCO bridges comprised the main synthetic challenge.
The diethynyl BCO was prepared using variants of known
(41) de Montigny, F.; Argouarch, G.; Lapinte, C. Synthesis 2006, (2), 293.
(42) Holtz, H. D.; Stock, L. M. J. Am. Chem. Soc. 1964, 86 (23), 5183.
(43) Baker, F. W.; Stock, L. M. J. Org. Chem. 1967, 32 (11), 3344.
(44) Kumar, K.; Wang, S. S.; Sukenik, C. N. J. Org. Chem. 1984, 49 (4),
665.
(45) Vongersdorff, J.; Kirste, B.; Niethammer, D.; Harrer, W.; Kurreck,
H. Magn. Reson. Chem. 1988, 26 (5), 416.
(46) Kilså, K.; Kajanus, J.; Mårtensson, J.; Albinsson, B. J. Phys. Chem.
B 1999, 103 (34), 7329.
(47) Caskey, D. C.; Wang, B.; Zheng, X. L.; Michl, J. Collect. Czech. Chem.
Commun. 2005, 70 (11), 1970.
(48) Chapman, N. B.; Sotheeswaran, S.; Toyne, K. J. J. Org. Chem. 1970,
35 (4), 917.
(49) Chu, Y. J.; Lynch, V.; Iverson, B. L. Tetrahedron 2006, 62 (23), 5536.
9
J. AM. CHEM. SOC. VOL. 130, NO. 24, 2008 7661

A R T I C L E S
Goldsmith et al.
Scheme 3a
a Reagents: (a) (1) SOCl2, Br2, MeOH, hν, (2) pyridine, 22%; (b) DIBALH, THF, 42%; (c) TBDMSCl, DMF, 66%; (d) p-benzoquinone, toluene, 83%;
(e) NaBH4, CeCl3 ·7H2O, MeOH, DCM, quant.; (f) POCl3, pyridine, 68%; (g) H2, Pd/C, EtOH, EtOAc, 90%; (h) TBAF, THF, 95%; (i) PCC, DCM, 73%;
(j) dimethyl-1-diazo-2-oxopropylphosphonate, K2CO3, MeOH, 87%; (k) (1) bromocyanoanthracene, TEA, CuI, Pd(PPh3)2Cl2, (2) N-(4-iodophenyl)pyrrolidine,
piperidine, CuI, Pd(PPh3)4, 5%.
Scheme 4a
a Reagents: (a) paraformaldehyde, HCl(g), dioxane, 50%; (b) NaOAc, HOAc, 78%; (c) toluene, C2H4 (2400 psi), quant.; (d) EtOH, KOH, 93%; (e) PCC,
DCM, 56%; (f) dimethyl-1-diazo-2-oxopropylphosphonate, K2CO3, MeOH, 92%; (g) (1) bromocyanoanthracene, TEA, CuI, Pd(PPh3)2Cl2, (2) N-(4-
iodophenyl)pyrrolidine, piperidine, CuI, Pd(PPh3)4, 7%.
products. Following procedures of Schmid50 and Lin,51 the
Diels-Alder adduct was reduced and rearomatized. The remain-
ing olefin was reduced, and the alcohols were deprotected.
Arbitrary functionality could be added at this point to the
bridgehead positions. To construct the alkynes we first oxidized
the alcohols to aldehydes. The Corey-Fuchs procedure failed
at this point to produce the desired alkyne. However, application
of Bestman’s reagent52 afforded the alkyne directly from the
aldehyde in good yield.
Synthesis of the DBCO system, Scheme 4, began with the
bis-chloromethylation of anthracene,53 followed by conversion
to the bis-acetoxy derivative.54 Previous attempts to produce
ethanoanthracenes relied on addition of maleic anhydride
followed by hydrolysis and bisdecarboxylation with lead
tetraacetate.55 This route is, however, inconsistent and low
yielding. The crucial step in our method is Diels-Alder addition
of ethene to 9,10-bis(acetoxymethyl)anthracene in toluene56 in
Figure 2. 1H NMR spectra of the aromatic regions of 1-4 at 300 K in
CDCl3. Red circles denote peaks from CA, blue triangles denote peaks from
PP, and green squares denote peaks from the bridge. Notice the downshifting
CA peaks.
asymmetrically substitute the BCO bridge. Other workers46,57
have reported the successful use of piperidine to couple electron-
rich aryl halides to bicyclooctyl or triptycenyl alkynes, and so
we attempted using it as the base and/or solvent. Optimization
of conditions allowed isolation of the title molecules after
purification via preparative HPLC.
a high pressure autoclave to quantitatively give the correspond-
ing ethanoanthracene derivative. The acetoxy groups were
hydrolyzed, and the alcohols were oxidized to aldehydes. The
Bestmann reagent was again applied to convert the aldehydes
directly to the alkyne.
Sonogashira coupling of the redox active groups to the BCO
bridges proved to be difficult, partially because of the reactivities
of the D and A aryl halides and partially because of the need to
The aromatic regions of the NMR spectra of 1-4 are shown
in Figure 2. As further evidence of the structure, one can see
the progressive downfield shift of the protons on the donor and
acceptor oriented toward the BCO bridge. As more phenyl
groups are fused to the bridge, these protons spend more time
caught in the deshielding ring current of the bridge phenyl
groups, resulting in this incremental shift downfield.
(50) Schmid, G. H.; Rabai, J. Synthesis 1988, (4), 332.
(51) Lin, C. T.; Wang, N. J.; Yeh, Y. L.; Chou, T. C. Tetrahedron 1995,
51 (10), 2907.
(52) Roth, G. J.; Liepold, B.; Muller, S. G.; Bestmann, H. J. Synthesis 2004,
(1), 59.
(53) Miller, M. W.; Amidon, R. W.; Tawney, P. O. J. Am. Chem. Soc.
1955, 77 (10), 2845.
(54) Badger, G. M.; Cook, J. W. J. Chem. Soc. 1939, 802.
(55) Chapman, N. B.; Sotheeswaran, S.; Toyne, K. J. Chem. Commun. 1965,
(11), 214.
(56) Meek, J. S.; Godefroi, V.; Wilcox, M. F.; Benson, W. R.; Clark, W. G.;
Tiedeman, T.; Evans, W. B. J. Org. Chem. 1961, 26 (11), 4281.
(57) Godinez, C. E.; Zepeda, G.; Garcia-Garibay, M. A. J. Am. Chem. Soc.
2002, 124 (17), 4701.
9
7662 J. AM. CHEM. SOC. VOL. 130, NO. 24, 2008

Similar Charge Transfer Rates through Benzo-annulation
A R T I C L E S
Figure 4. Steady-state UV-vis spectra of 1-5 and 7 in toluene at 295 K.
Figure 3. Structures of model compounds.
Table 1. Measured Redox Potentials in V versus SCE and
Relevant Electron Transfer Parameters for 1-4 in eV
‡
Eox
Ered
∆GCS
λI
λO
∆GCS
0.74
-1.27
Toluene
MTHF
-0.02
-0.57
0.21
0.21
0.05
0.72
0.055
0.034
Energy Levels. The donor and acceptor were chosen to meet
specific redox and spectroscopic requirements and have been
used in previous photoinduced electron transfer studies.58
Compounds 5 and 6 were prepared as appropriate models, Figure
3. Cyclic voltammetry was performed, and the results are
summarized in Table 1. These values were used to estimate the
energy of the radical pair state by employing Weller’s expres-
sion,59 which relies on the Born dielectric continuum model.
The free energy change for charge separation, ∆GCS, from the
locally excited CA group can then be calculated. To calculate
internal reorganization energies for the donor and acceptor, λI,
we first geometry optimized the neutral species and the
appropriate ions of 5 and 6. Taking the difference in self-
consistent field (SCF) energy of the ions calculated in the neutral
and ionic geometries yielded λI. The solvent reorganization
energy, λS, was calculated from Marcus’ formula60 assuming a
dielectric continuum. Finally, using the classical Marcus expres-
sion,60 the activation barrier for charge separation can be
calculated. Results for toluene and MTHF are shown in Table
1, and sample calculations are presented in the Supporting
Information.
Spectroscopy. Ground state absorption spectra of 1-5 and
another model compound, 7 (Figure 3), are shown in Figure 4.
The CA acceptor dominates the absorption in this region,
although the progressive red shift of the bridge absorption can
also be seen. Comparison of 1-4 with 5 shows only minimal
deviations, showing that the CA acceptor is effectively elec-
tronically isolated from the rest of the molecule. Steady-state
fluorescence (not shown) is also similar and dominated by the
CA acceptor. In contrast, the absorption spectrum of 7 is very
different, indicative of significant communication between the
donor and acceptor.
Figure 5. Transient absorption spectra of 2 at 295 K in toluene.
CA-•. The transient absorption spectra are dominated by
absorption of 1*CA at early times and CA-• at later times, as
shown in Figure 5. The observed CA-• absorption is qualita-
tively similar to the electrochemically generated anion of a
related substituted CA compound.58 Transient absorption spectra
of 5 in toluene shows only 1*CA, while 1-4 and 7 show the
distinct features of CA-•. Transient absorption spectra of 1-4
show different λmax values for CA-• in MTHF, while all the
spectra are very similar in toluene, Figure 6.
Charge separation (CS) and recombination (CR) dynamics
can be examined by measuring the rise and the fall of the CA-•
feature. Representative kinetics for 1-4 are shown in Figure 7.
while time constants for CS and CR are summarized in Table
2. The CS and CR time constants for 1-4 in toluene at 295 K
are very similar. CS in 7 is comparable to or faster than the
instrument response function (τIRF ) 0.18 ps). The CS of 2 and
3 in MTHF at 295 K is marginally slower than that of 1 and 4.
CS times in glassy MTHF at 80 K are faster than those at room
temperature, with CS in 1 occurring marginally faster than that
in 2-4. CR for 1-4 in MTHF at 295 K is too slow to be
accurately measured using our ultrafast transient absorption
apparatus but too fast to be measured with our nanosecond
transient absorption apparatus (τIRF ) 7 ns), although the time
constants are all in the range 2-4 ns.
Discussion
Photoexcitation of CA at either 416 or 430 nm results in
Molecules 1-4 were designed to show the effect of an
increasing number of π-pathways on the rate of charge transfer.
Implicit in our design was the assumption that any π-pathway
would significantly accelerate charge transfer over a σ-pathway.
However, the essentially equal CS and CR rate constants of
1-4 show that the π-system has no overall effect, contrary to
efficient charge transfer to form the radical ion pair, PP+•-B-
(58) Elangovan, A.; Kao, K. M.; Yang, S. W.; Chen, Y. L.; Ho, T. I.; Su,
Y. L. O. J. Org. Chem. 2005, 70 (11), 4460.
(59) Weller, A. Z. Phys. Chem. 1982, 130 (2), 129.
(60) Marcus, R. A.; Sutin, N. Biochim. Biophys. Acta 1985, 811 (3), 265.
9
J. AM. CHEM. SOC. VOL. 130, NO. 24, 2008 7663

A R T I C L E S
Goldsmith et al.
Figure 6. Transient absorption spectra of 1-4 at 295 K in (a) toluene at 1029 ps and (b) in MTHF at 239 ps.
Figure 7. Transient absorption kinetics of 1-4 measured in (a) toluene at 643 nm at 295 K (CS); (b) toluene at 655 nm at 295 K (CR); (c) MTHF at 613
nm at 295 K (CS); and (d) MTHF at 716 nm at 80 K (CS).
Table 2. Electron transfer time constants of 1-4
τCS(toluene)
(ps)
τCR(toluene)
(ns)
τCS(MTHF,300K)
(ps)
τCS(MTHF,80K)
(ps)
1
2
3
4
130 ( 12
145 ( 17
133 ( 5
48 ( 5
45 ( 5
54 ( 4
55 ( 4
27 ( 3
44 ( 5
33 ( 3
22 ( 4
2.5 ( 1.6
6.5 ( 1.0
3.9 ( 1.9
6.7 ( 2.7
141 ( 10
our expectations based on McConnell’s expression. Clearly, the
relative roles of the π- and σ-systems need to be re-examined.
Contribution of the Bridge π-System. The charge separation
rate of 7 was significantly faster than that in 1-4, consistent
with our expectations from eq 1 and previous work,21 showing
that a conjugated π-system does provide greater electronic
communication than a σ-system in an equivalent geometry. To
understand why our nonconjugated π-pathways fail to accelerate
kCT, we estimated the respective overlaps of the π system of
the chromophores with the π- and σ-systems on the bridge,
focusing on nearest and next-nearest neighbor interactions, as
Figure 8. Relevant orbitals for overlap with the bridge σ- and π-systems.
S is the spatial overlap, and F is the appropriate Fock matrix element.
shown in Figure 8. We examine the overlap on a per “rung”
basis (i.e., though the σ-system has three equivalent σ pathways,
we will focus on one at time where each rung is a set of two
methylene carbons) as well as focus on the impact of one
additional phenyl ring at a time, as seen in Figure 8. The
9
7664 J. AM. CHEM. SOC. VOL. 130, NO. 24, 2008

Similar Charge Transfer Rates through Benzo-annulation
A R T I C L E S
proximal alkyne pπ orbital will be considered the only point of
communication between the chromophores and the bridge.
Hydrogen-like angular functions were used61 with two types
of radial functions. We used the radial function of the slowest
decaying carbon Gaussian-type orbital (GTO) for the 6-31G*
basis set. This primitive orbital had a decay constant of 0.48
Å-2. We also used a simple exponentially decaying radial
function with a decay constant of 1.5 Å-1 which was based on
the calculated ꢀ distance parameter for the through-space
distance dependence of electron transfer, 3 Å .
-1 62 Full calcula-
tion details including relevant bond angles are given in the
Supporting Information. The overlap between the alkyne pπ
orbital and the bridge σ-system is only two to three times larger
than the overlap with the marginally more distant π-system,
Figure 8. We had hoped that this disparity would be overcome
by the more favorable energy resonance condition resulting from
the denominator of eq 1 due to the communication being
between two π-systems.
Figure 9. Application of Walton and Adcock’s orbital overlap model.67
clo[2.2.2]octatriene (barrelene), which showed incrementally
dropping σ-system orbital energies from -9.7 to -11.3 eV.66
This energy drop is central to the explanation of an important
experiment relevant to this work by Walton, Adcock and co-
workers, who examined the hyperfine coupling between bridge-
head radicals and functionality at the opposite bridgehead in
bicyclo[2.2.2]octane and triptycene derivatives.67 They found
significant hyperfine coupling through the bicyclo[2.2.2]octane
system to a variety of substituents at the other bridgehead but
found no coupling in the triptycene system for any of the
substituents. Frequently, it is assumed that hyperfine couplings
are communicated to a particular atom via the same superex-
change type interactions that mediate electron transfer because
the magnitude of the hyperfine coupling is proportional to the
local population of the unpaired electron at that nucleus.
Consequently, any model proposed for electronic interactions
in our system must also be able to accommodate their results.
Our expectation was that the π-pathways would increase
communication between bridgehead anchored donor and ac-
ceptor groups. Our experimental result is that the π-pathways
have a negligible effect. However, the Walton, Adcock, et al.
experiments suggest that the π-pathways decrease communica-
tion. Coupling is suggested to be mediated via overlap between
a series of antiperiplanar sp3 hybridized orbitals. Such an
orientation is optimal for maximum overlap (Cieplak’s model).68,69
To explain their result, the authors67 suggest the dominance
of an antiperiplanar arrangement of σ-orbitals in which the
dominant communication conduit is the σ-bond oriented parallel
to the bridgehead radical orbital and the C-H bond of the
opposite bridgehead, Figure 9a. The direct through-space
interaction between the bridgeheads has been shown to be
minimal,70 while this “skipped” through-bond mechanism (as
opposed to going through consecutive bonds) is supported by
measurements of hyperfine couplings in related adamantane
radicals, where the coupling constants to two geminal hydrogens
would be equivalent in the case of consecutive through-bond
coupling, but are significantly different due to the presence of
antiperiplanar overlap.71 As mentioned above, fusing the phenyl
rings to the BCO changes the hybridization of this bond,
removing an energetic resonance condition and decreasing the
We also undertook Natural Bond Orbital63 (NBO) analysis
to probe the significance of the interaction between the more
distal π orbitals. By expressing the Fock matrix in the local
NBO basis, we can examine the appropriate off-diagonal
elements to probe these specific microscopic interactions. Such
an approach has been used to great advantage in pathways type
analyses of transport in organic molecules.6,10–12 In addition,
direct examination of the matrix element avoids the use of a
Mulliken “magic rule” type expression64 where the matrix
element is assumed to be directly proportional to the orbital
overlap, eq 2,
Vij ) ψi|H|ψj Sij ) ψi|ψj
(2)
where the Ψi refer to the component wave functions.
In contrast to the simple analysis above, the NBO analysis
shows a nearly vanishing matrix element between the alkyne
and bridge π-systems. Clearly, the application of Mulliken’s
“magic rule” proportionality is shown to be an inadequate
description for our system and suggests caution in the blanket
application of eq 2. The bridge π-system is thus effectiVely
isolated from the π-systems of the chromophores. Additionally,
the isolation of the π-systems is corroborated by measurements
of hyperfine couplings of the phosphorus centered radical of
X-irradiated 9-phosphinotriptycene.65 The radical shows no
hyperfine coupling to any part of the triptycene and only dipolar
coupling to the phenyl hydrogens proximal to the phosphorus.
Contribution of the Bridge σ-System. When examining the
significant electronic changes occurring upon benzo-annulation,
it is tempting to focus entirely on the π-system. However, the
σ-system should also undergo significant electronic changes as
the hybridization of the bonding orbitals shifts from sp3 to sp2.
Specifically, the increasing s character of the orbitals should
result in a lower orbital energy. This trend has been seen
experimentally via photoelectron spectroscopy in the study of
BCO, bicyclo[2.2.2]octene, bicyclo[2.2.2]octadiene, and bicy-
(61) DeKock, R. L.; Gray, H. B. Chemical Structure and Bonding;
University Science Books: Sausalito, CA, 1989.
(66) Haselbach, E.; Heilbronner, E.; Schroder, G. HelV. Chim. Acta 1971,
54 (1), 153.
(62) Paddon-Row, M. N.; Jordan, K. D. Modern Models of Bonding and
Delocalization; VCH, Weinheim, New York, 1988; Vol. 6.
(63) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. ReV. 1988, 88 (6),
899.
(67) Binmore, G. T.; Walton, J. C.; Adcock, W.; Clark, C. I.; Krstic, A. R.
Magn. Reson. Chem. 1995, 33, S53.
(68) Cieplak, A. S. J. Am. Chem. Soc. 1981, 103 (15), 4540.
(69) Tomoda, S. Chem. ReV. 1999, 99 (5), 1243.
(70) Hoffman, R. Acc. Chem. Res. 1970, 4 (1), 1.
(71) Krusic, P. J.; Schleyer, P. V.; Rettig, T. A. J. Am. Chem. Soc. 1972,
94 (3), 995.
(64) Mulliken, R. S. J. Phys. Chem. 1952, 56, 295.
(65) Ramakrishnan, G.; Jouaiti, A.; Geoffroy, M.; Bernardinelli, G. J. Phys.
Chem. 1996, 100 (26), 10861.
9
J. AM. CHEM. SOC. VOL. 130, NO. 24, 2008 7665

A R T I C L E S
Goldsmith et al.
Table 3. Parameters from NBO Analysisa
bond mixing over consecutive bond mixing in the antibonding
orbitals (not shown), consistent with NBO analysis of BCO
anions by Liang and Newton.10 Charge transfer in our system
is likely proceeding via a hole transfer mechanism because the
electron acceptor is first locally excited, so we will focus on
the bonding orbitals. However, mixing through the bonding
orbitals shows a complex network of destructive interference
within each methylene rung, again consistent with previous
work.10 As a consequence of the destructive interference the
NBO analysis predicts that the majority of through-bond
communication involves NBO 3. Thus, NBO analysis suggests
significant contributions from NBO 3 to the total Veff and an
effect from the significant perturbation of that bond accompany-
ing introduction of a phenyl group, and so does not support the
models in Figure 9.
no phenyl group
with phenyl group
F(1,1)
F(2,2)
F(3,3)
F(1,2)
F(1,3)
F(1,4)
F(1,5)
F(2,3)
F(2,4)
-7317
-15640
-15667
-898
264
-41
2.7
-3237
-816
-7371
-16402
-18360
-816
218
-33
2.7
-3634
-686
Direct
1 Step
Re-examining Walton and Adcock’s result, we calculated the
hyperfine coupling at the bridgehead hydrogen and found the
Veff 1,5
2.7
2.7
expected disparity in magnitude between the BCO (hfcBCO-H
)
Veff 1,2,5
Veff 1,3,5
Veff 1,4,5
4.4
8.3
4.4
2.9
4.3
2.9
3.2G) and triptycene radicals (hfcTrip-H ) -0.18G). However,
hyperfine coupling constants to the adjacent bridgehead carbon,
which were not observed experimentally, did not show a similar
disparity in magnitude (hfcTrip-C ) 0.75G, hfcTrip-H ) -0.74G).
If delocalization of the radical via antiperiplanar overlap was
the dominant interaction, one would expect the magnitude of
spin density at the carbon and hydrogen to track each other.
Delocalization of the unpaired electron is only one mechanism
to attain nonvanishing spin density.72 In the study of [1.1.1]bi-
cyclopentane (staffane) σ radicals, two mechanisms of spin
propagation were identified: spin delocalization and spin po-
larization.73 NBO deletion analysis showed that spin polarization
was the dominant mechanism. It is expected to be even more
dominant in a BCO system where through-bond and through-
space delocalization will be even weaker due to the large
distance between bridgeheads. Hyperfine coupling from spin
polarization is extremely sensitive to the hybridization of
surrounding atoms,72 with the benzo-annulation consequently
having a large effect on the polarization. As a result, we find
no conflict with Walton and Adcock’s suggestion that the change
in hybridization reduces the spin density, only that it is a result
of the change in polarization rather than a reduction in
delocalization. Most importantly, since it is this delocalization
that is most relevant when comparing hyperfine couplings and
Veff for charge transfer, these results are likely not in conflict
2 Step
3 Step
Veff 1,2,3,5
Veff 1,2,4,5
Veff 1,3,4,5
11
-9.5
11
6.5
-5.6
6.5
Veff 1,2,3,4,5
15
47
9.8
30
total
a Matrix elements are given in meV. The i,j refer to the numbered
orbitals above. F(i,j) refer to the elements of the Fock matrix. Veff(i,j,...)
are calculated using eq 1. The model system is symmetric across NBO 3.
with our observation of unchanging kCT
.
Nonetheless, the NBO analysis and photoelectron spectra
mentioned above strongly suggest consequences of the changing
hybridization for kCT. From a different perspective, one can
simply view 1-4 as three-site π-σ-π-systems with the lowest
σ-orbital energy Eσ ) -9.7 eV for BCO in 1 and Eσ ) -11.3
eV for triptycene in 4 estimated from the barrelene data.66
Taking Eπ ) -7.8 eV from photoelectron spectra of 9-cyanoan-
thracene,74 we can use eq 1 to estimate the difference in Veff
going from 1 to 4 to be a reduction by a factor of 0.54,
corresponding to a reduction in kCT of 0.29. This reduction can
be relaxed significantly by increasing the energy gap between
the π- and σ-systems, and so one origin of our static rates could
be underestimation of this gap. Alternatively, eq 1, whose
Figure 10. Steady-state UV-vis of 1 at a variety of temperatures in MTHF.
effective mixing, consistent with McConnell’s relation, eq 1.
This explanation can be conveniently adapted to suit our system
geometry, as shown in Figure 9b. Analogously, the p orbitals
of the alkyne should be in good periplanar overlap with the σ
bonds originating at the bridgehead, which are in turn in good
periplanar overlap with the σ bond at the opposite bridgehead,
and sequentially the p orbital at the opposite alkyne (NBO’s
1-2-4-5 in Table 3). Because this model suggests minimal
importance for the central σ orbital (NBO 3 in Table 3), a change
in hybridization of this orbital would be predicted to have a
minimal effect on Veff.
NBO analysis for the bonding orbitals of the bridge σ-system
is summarized in Table 3. The analysis captures the effect of
the changing hybridization that accompanies the addition of the
phenyl ring by showing a decrease in bonding orbital energy.
NBO analysis also captures the dominance of skipped through-
(72) Wertz, J. E.; Bolton, J. R. Electron Spin Resonance: Elementary Theory
and Practical Applications; McGraw-Hill: New York, 1972.
(73) McKinley, A. J.; Ibrahim, P. N.; Balaji, V.; Michl, J. J. Am. Chem.
Soc. 1992, 114 (26), 10631.
(74) Klasinc, L.; Kovac, B.; Gusten, H. Pure Appl. Chem. 1983, 55 (2),
289.
9
7666 J. AM. CHEM. SOC. VOL. 130, NO. 24, 2008

Similar Charge Transfer Rates through Benzo-annulation
A R T I C L E S
Figure 11. Charge transfer trajectories for a model of 1+ at different effective temperatures. (a) Molecular dynamics snapshot with zero point energy (0 K)
in each of the modes and the corresponding trajectory (b) showing charge shift. (c) A molecular dynamics snapshot after equilibration at 300 K with (d) no
charge shift in the trajectory.
validity requires a high Ei/Vij ratio, can also be significantly
overestimating the reduction. However use of the semiclassical
Marcus equation in the nonadiabatic limit,
of summing pathways coherently.27,28,77 However, these fluctua-
tions must occur on a sufficiently fast time scale to influence
the additivity of the different pathways. A convenient method
of estimating this time threshold is by calculating the
Landauer-Bu¨ttiker tunneling time,78 a concept adapted for
Donor-Bridge-Acceptor molecules by Nitzan and co-work-
ers79 and shown in eq 4,
2π
1
-∆G‡
kCS
)
V2eff
exp
(3)
(
)
p
4πλk T
kBT
ꢀ
B
with ∆G‡ ) (λ+∆GRP)2/4λ and the parameters in Tables 1 and
2 suggest a Veff in the range of several meV. Using eq 1 with
this Veff and the above Ei suggests a microscopic Vij in the range
of tens of meV, a range where eq 1 should be valid.
pN
τtun
)
(4)
∆E
where N is the number of bridge sites and ∆E is the energetic
gap between the bridge and the D and A. Conceptually, this is
the time scale at which bridge fluctuations must occur in order
to influence the charge carrier using the bridge. When the
transfer occurs via superexchange (large gap), these fluctuations
must be incredibly fast in order to influence the charge carrier.
This is consistent with the charge carrier never actually
occupying the bridge and being consequently decoupled from
its fluctuation. On the other hand, charge transfer operating with
small gaps or in the hopping regime should be responsive to
relatively slow fluctuations, since the actual occupation of the
bridge makes the charge carrier more sensitive to fluctuations.
With a gap of 1.9 eV (as estimated from the ionization potentials
of BCO, 9.7 eV,66 and CA, 7.8 eV74), the corresponding
tunneling times of 0.3 fs suggest that if any interference did
The NBO analysis predicts a significant change in Veff upon
benzo-annulation as a result of the changing hybridization and
many interfering pathways within each rung, while our experi-
mental results suggest a single dominant pathway that is
unaffected by the change in hybridization. One way to achieve
a dominant pathway relies on the interference between different
orbitals within the σ-system (whether it be constructive or
destructive) being destroyed by electronic dephasing, leaving
one dominant pathway. Contributions from multiple spatial
pathways require well-defined phase relationships of the charge
carrier wave function to be maintained at all sites along the
multiple pathways. The fidelity of this phase relationship can
be conveniently expressed as the off-diagonal element (coher-
ence) of the density matrix. Fluctuations in site energies (as
well as intersite coupling) that exist at finite temperature can
destroy these phase relationships75,76 and spoil the legitimacy
(77) Skourtis, S. S.; Waldeck, D. H.; Beratan, D. N. J. Phys. Chem. B
2004, 108 (40), 15511.
(78) Büttiker, M.; Landauer, R. Phys. ReV. Lett. 1982, 49 (23), 1739.
(79) Nitzan, A.; Jortner, J.; Wilkie, J.; Burin, A. L.; Ratner, M. A. J. Phys.
Chem. B 2000, 104 (24), 5661.
(75) Skinner, J. L.; Hsu, D. J. Phys. Chem. 1986, 90 (21), 4931.
(76) Onuchic, J. N.; Wolynes, P. G. J. Phys. Chem. 1988, 92 (23), 6495.
9
J. AM. CHEM. SOC. VOL. 130, NO. 24, 2008 7667

A R T I C L E S
Goldsmith et al.
exist, it would survive the electronic dephasing. Thus, dephasing
should not erase interference in this case to leave a single
dominant pathway.
temperature may reduce the electronic dephasing rate enough
to observe properties that rely on coherent summation of the
charge carrier wave function over multiple spatially distinct parts
of the molecule.27,28 Despite frequently being referred to as rigid
spacers, substituted alkynes can contort significantly at finite
temperature.84 Flexibility in D-B-A molecules can lead to the
electronic properties being dominated by conformations other
than the equilibrium one. This scenario was seen in molecules
where charge transfer is symmetry-forbidden85,86 and is el-
egantly captured in the simulations of Jones et al.87
The ground-state absorption spectra of 1 in MTHF are shown
in Figure 10 at 300, 160, and 80 K. The sharpening of the
features reveals two details about the system. First, though
cyanoanthracene derivatives have been known to aggre-
gate,58,88 no aggregation occurs here as the absorption spectra
show no evidence for the requisite exciton coupling89 that
accompanies aggregation.90 Second, the reduction in inhomo-
geneous broadening suggests that the ensemble of molecules is
adopting fewer different conformations. This is consistent with
depopulation of excited conformational modes before solidifica-
tion. The sharpened features imply that the conformations under
study in the glass are closer to the equilibrium conformation
than those in the solution phase.
Upon freezing, molecules 1-4 show monoexponential CS
kinetics with significantly faster CS rates, although the reduced
signal-to-noise may obscure the presence of additional compo-
nents. Such behavior is reminiscent of the bacterial photosyn-
thetic reaction center.91,92 In that case the rate acceleration was
explained as higher vibrations taking the system away from the
bottom of the vibronic well, the optimal position for activa-
tionless charge transfer.93 However, nonvanishing activation
energies, Table 1, suggest this explanation is not valid in our
system. More likely, the rise in CS rate is due to a conforma-
tional gating effect,36,85,86,94 whereby the deviations from
equilibrium conformation result in a reduction of Veff. These
less efficient conformations are accessed more frequently at 300
K than at 80 K. To test this hypothesis we followed the example
of Jones and co-workers95 and performed molecular dynamics
coupled with semiempirical calculations to follow the charge
transfer trajectories. Our simulated molecule is a model of 1
with a +1 charge, Figure 11. When only the zero-point energy
was injected into the normal modes of the computational model
system (T ) 0 K), only geometries close to the equilibrium
geometry were observed, Figure 11a, and a charge shift occurred
within several hundred femtoseconds of the simulation, Figure
11b. In contrast, when additional energy was put into the system
The remaining alternative explanation simply restates that
NBO analysis, though significantly informative in relating
experimental observables to physical organic concepts,63 is
ultimately qualitative in nature. The importance of looking
beyond nearest neighbors has been consistently shown through
NBO calculations,10,11,62 but the dominance of a single interac-
tion may be overlooked by the NBO analysis. In this case, that
single interaction would necessarily not involve the central σ
bond (NBO 3), and would preserve the model in Figure 9.
MTHF Data at 300 K. Greater insight into the charge transfer
mechanism may also be obtained by shifting to different solvents
and temperatures. CS rates in MTHF show an increase over
the toluene rates. This increase is consistent with the predictions
of Marcus theory following the activation barriers in Table 1.
In contrast to the charge separation rates in toluene, the rates
measured in MTHF show a small but experimentally reproduc-
ible difference in rates with 1 and 4 being slightly faster than
2 and 3. While we had initially viewed this difference in rates
as a consequence of the altered σ-system, a more likely
explanation can be drawn from the transient absorption spectra,
Figure 6. While the spectra show only marginal differences in
toluene, 1-4 show reproducible and relatively significant
differences in MTHF. Since the ground-state absorption shows
no difference among 1-4, and the rates measured in toluene
are essentially equivalent, the changes in rates and spectra in
MTHF are likely attributed to changes in solvation of the ions
in the radical pair.
Achieving the necessary solvent polarization is central to the
Marcus theory of electron transfer,60 comprising the critical
reaction coordinate in the original formulation of the electron
transfer reaction. The first and second solvent shells provide
the critical stabilization for the shifting charges.80,81 The
additional bridge phenyl groups in 2-4 are located within the
range of the first and second shells, and so we speculate that
there will be a consequent disruption of the solvation shell. Such
a steric disturbance of the solvent packing has previously been
suggested to influence redox behavior.82 Any differential
solvation effect would be magnified in MTHF over toluene
because of MTHF’s higher dipole moment. That the rate does
not vary monotonically with additional phenyl groups and the
observation that it is the two 3-fold symmetric species that show
the fastest rates suggest symmetry has a role maintaining optimal
solvation. These systems would provide an interesting object
of study for solvent molecular dynamics simulations. A similar
nonmonotonic trend also suggesting the importance of symmetry
was seen in the relative rates of bromine abstraction at the
bridgehead position of a similar series of molecules.83
(84) Kelly, T. R. Acc. Chem. Res. 2001, 34 (6), 514.
(85) Balzani, V.; Barigelletti, F.; Belser, P.; Bernhard, S.; De Cola, L.;
Flamigni, L. J. Phys. Chem. 1996, 100 (42), 16786.
(86) Zeng, Y.; Zimmt, M. B. J. Am. Chem. Soc. 1991, 113 (13), 5107.
(87) Jones, G. A.; Paddon-Row, M. N.; Carpenter, B. K.; Piotrowiak, P. J.
Phys. Chem. A 2002, 106 (19), 5011.
MTHF Data at 80 K. Evaluation of charge separation rates
in a frozen glass serves two purposes. By reducing the
population of high vibrational levels one can limit the confor-
mational space explored by the molecule and learn about which
conformations favor charge transfer. In addition, lowering the
(88) Morsi, S. E.; Carr, D.; El-Bayoumi, M. A. Chem. Phys. Lett. 1978,
58 (4), 571.
(89) Kasha, M.; Rawls, H. R.; El-Bayoumi, M. A. Pure Appl. Chem. 1965,
11, 371.
(90) Acharya, S.; Bhattacharjee, D.; Talapatra, G. B. Chem. Phys. Lett.
2002, 352 (5-6), 429.
(80) Barthel, E. R.; Martini, I. B.; Schwartz, B. J. J. Phys. Chem. B 2001,
105 (49), 12230.
(91) Fleming, G. R.; Martin, J. L.; Breton, J. Nature 1988, 333 (6169),
190.
(81) Kuharski, R. A.; Bader, J. S.; Chandler, D.; Sprik, M.; Klein, M. L.;
Impey, R. W. J. Chem. Phys. 1988, 89 (5), 3248.
(82) Meier, M.; Sun, J.; Wishart, J. F.; van Eldik, R. Inorg. Chem. 1996,
35 (6), 1564.
(92) Hoff, A. J.; Deisenhofer, J. Phys. Rep. 1997, 287 (1-2), 2.
(93) Bixon, M.; Jortner, J. J. Phys. Chem. 1986, 90 (16), 3795.
(94) Davis, W. B.; Ratner, M. A.; Wasielewski, M. R. J. Am. Chem. Soc.
2001, 123 (32), 7877.
(83) Tabushi, I.; Yoshida, Z.; Aoyama, Y. Bull. Chem. Soc. Jpn. 1974, 47
(12), 3079.
(95) Jones, G. A.; Carpenter, B. K.; Paddon-Row, M. N. J. Am. Chem.
Soc. 1998, 120 (22), 5499.
9
7668 J. AM. CHEM. SOC. VOL. 130, NO. 24, 2008

Similar Charge Transfer Rates through Benzo-annulation
A R T I C L E S
(T ) 300 K), the geometry deviated significantly from the
equilibrium position, Figure 11c, and a charge shift was never
observed, Figure 11d. These simulations confirm our hypothesis
regarding the increased efficacy of the equilibrium conformation.
An additional report detailing specific per-mode contributions
is in progress.
The charge separation rate of 1 is seen to be marginally faster
than 2-4, although differences among 2-4 are difficult to probe
due to degradation of the sample and the consequent size of
the error bars. The relative steric bulk of 2-4 compared to 1
may also play a role in how the molecule interacts with the
MTHF matrix. Since the trend here is loosely similar to that
observed in solution and steric effects have been shown to be
significant, the lower electronic dephasing rate due to the lower
temperature is seen to be unimportant.
electronic dephasing in effectively preventing the coherent
addition of multiple σ-pathways. Charge separation rates in
MTHF showed some differences between members of the series
but were likely caused by differences in solvation. Charge
separation rates increased at lower temperatures because the
deviations from the equilibrium conformation experienced more
severely at high temperatures decrease the electronic com-
munication. This behavior was captured in molecular dynamics
simulations coupled to semiempirical electronic structure
calculations.
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
(MRW), and the MURI program of the DOD and the Chemistry
and International Divisions of the NSF (M.A.R.). The work at Penn
State was supported by the Chemical Sciences, Geosciences, and
Biosciences Division, Office of Basic Energy Sciences, DOE. We
thank Abraham Nitzan, Masanori Iimura, Barry Carpenter, Marshall
Newton, Larry Curtiss, Josef Michl, Qixi Mi, Garth Jones, and
Maxim Ovchinnikov for very helpful conversations. R.H.G. thanks
the Link Foundation and the Dan David Foundation for fellowships.
J.V.-W. thanks the NSF for a fellowship.
Conclusions
We have prepared a series of molecules to test the effect of
additional π-pathways fused to a σ-scaffold on mediating
electron transfer between two chromophores. Surprisingly, the
π-pathways were shown to have no role in the electron transfer,
and this situation was accurately captured by NBO analysis.
Equally surprisingly, we found that the changes in the σ-system
as a result of shifting hybridization also had no effect, despite
the predictions of photoelectron spectra. We were able to suggest
a series of relevant bonding orbitals that may be uninfluenced
by the changing hybridization. However, we were unable to
corroborate their dominance via NBO analysis possibly because
of its own qualitative nature or because of the importance of
Supporting Information Available: Characterization of new
molecules, synthetic procedures, sample calculations, cyclic
voltammograms, full ref 37. This material is available free of
charge via the Internet at http://pubs.acs.org.
JA8004623
9
J. AM. CHEM. SOC. VOL. 130, NO. 24, 2008 7669