Electronic Couplings in Organic Mixed-Valence Compounds: The Contribution of Photoelectron Spectroscopy

Article abstract of DOI:10.1021/ja039263u

We show that the electronic coupling in strongly coupled organic mixed-valence systems can be effectively probed by means of gas-phase ultraviolet photoelectron spectroscopy (UPS). Taking six diamines as examples, the UPS estimates for the electronic coup

Full text of DOI:10.1021/ja039263u

Published on Web 02/07/2004
Electronic Couplings in Organic Mixed-Valence Compounds:
The Contribution of Photoelectron Spectroscopy
Veaceslav Coropceanu,*,† Nadine E. Gruhn,‡ Stephen Barlow,†,‡
Christoph Lambert,§ Jason C. Durivage,‡ Tonja G. Bill,‡ Gilbert No¨ll,§
Seth R. Marder,†,‡ and Jean-Luc Bre´das*,†
Contribution from the School of Chemistry and Biochemistry, Georgia Institute of Technology,
Atlanta, Georgia 30332-0400, Department of Chemistry, The UniVersity of Arizona,
Tucson, Arizona 85721-0041, and Institut fu¨r Organische Chemie, Bayerische
Julius-Maximilians-UniVersita¨t Wu¨rzburg, Am Hubland, D-97074 Wu¨rzburg, Germany
Received October 27, 2003; E-mail: jean-luc.bredas@chemistry.gatech.edu; veaceslav.coropceanu@chemistry.gatech.edu
Abstract: We show that the electronic coupling in strongly coupled organic mixed-valence systems can
be effectively probed by means of gas-phase ultraviolet photoelectron spectroscopy (UPS). Taking six
diamines as examples, the UPS estimates for the electronic couplings Hab are compared with the
corresponding values determined from the intervalence charge-transfer absorption bands and from electronic
structure calculations. Similar trends are observed for the Hab values estimated from UPS and optical spectra;
this provides support for the applicability of Hush theory to strongly coupled organic mixed-valence systems.
The UPS electronic couplings are found to be somewhat smaller than those from optical spectroscopy,
which is attributed to the role of vibronic coupling to symmetrical modes; when corrected for this vibronic
coupling, the UPS Hab estimates confirm that triarylamine-based mixed-valence systems are close to the
class-II/class-III borderline.
for example, bis(diarylamino)biphenyl9 derivatives are com-
monly used as hole transport agents in organic light-emitting
diodes where the charge carrier is the corresponding mixed-
valence cation. Switching devices based on organic mixed-
valence systems have also been proposed.10 The properties of
MV systems strongly depend on the extent of electronic coupling
between charge-bearing subunits. The magnitude of this interac-
tion is defined by the matrix element Hab ) Ψa|H|Ψb , where
H is the molecular Hamiltonian and Ψa and Ψb are the wave
functions of the diabatic states, corresponding in dimeric MV
systems to the two possible valence structures Ma+-BR-Mb
Introduction
Organic mixed-valence (MV) systems are receiving increasing
attention because of the insight they provide into the under-
standing of electron-transfer phenomena.1-8 They are also of
interest due to their relevance to organic electronic materials;
† Georgia Institute of Technology.
‡ The University of Arizona.
§ Bayerische Julius-Maxmilians-Universita¨t Wu¨rzburg.
(1) (a) Lahlil, K.; Moradpour, A.; Bowlas, C.; Menou, F.; Cassoux, P.;
Bonvoisin, J.; Launay, J.-P.; Dive, G.; Dehareng, D. J. Am. Chem. Soc.
1995, 117, 9995. (b) Ruiz-Molina, D.; Sedo, J.; Rovira, C.; Veciana, J.
Handb. AdV. Electron. Photon. Mater. DeVices 2001, 3, 303. (c) Rovira,
C.; Ruiz-Molina, D.; Elsner, O.; Vidal-Gancedo, J.; Bonvoisin, J.; Launay,
J. P.; Veciana, J. Chem. Eur. J. 2001, 7, 240. (d) Gautier, N.; Dumur, F.;
Lloveras, V.; Vidal-Gancedo, J.; Veciana, J.; Rovira, C.; Hudhomme, P.
Angew. Chem., Int. Ed. 2003, 42, 2765.
+
or Ma-BR-Mb (with M denoting a redox site and BR the
bridge).
The electronic couplings are usually estimated from the
analysis of the intervalence charge-transfer (IV-CT) transition.
A simple model that has been extensively verified for transition-
metal MV compounds and is now increasingly applied to
organic MV systems was elaborated by Hush.11 Using a
semiclassical formalism, Hush showed that the electronic
coupling is related to the maximum, Eop, and transition dipole
moment, µab, of the IV-CT band as11
(2) (a) Lindeman, S. V.; Rosokha, S. V.; Sun, D.; Kochi, J. K. J. Am. Chem.
Soc. 2002, 124, 843. (b) Rosokha, S. V.; Sun, D. L.; Kochi, J. K. J. Phys.
Chem. A 2002, 106, 2283.
(3) (a) Bailey, S. E.; Zink, J. I.; Nelsen, S. F. J. Am. Chem. Soc. 2003, 125,
5939. (b) Nelsen, S. F.; Ismagilov, R. F.; Trieber, D. A. Science 1997,
278, 846. (c) Nelsen, S. F.; Ismagilov, R. F.; Powell, D. R. J. Am. Chem.
Soc. 1996, 118, 6313. (d) Nelsen, S. F.; Ismagilov, R. F.; Powell, D. R. J.
Am. Chem. Soc. 1997, 119, 10213. (e) Nelsen, S. F.; Ismagilov, R. F.;
Powell, D. R. J. Am. Chem. Soc. 1998, 120, 1924. (f) Nelsen, S. F.; Tran,
H. Q.; Nagy, M. A. J. Am. Chem. Soc. 1998, 120, 298. (g) Nelsen, S. F.;
Konradsson, A. E.; Weaver, M. N.; Telo, J. P. J. Am. Chem. Soc. 2003,
125, 12493.
(4) Mayor, M.; Buschel, M.; Fromm, K. M.; Lehn, J. M.; Daub, J. Chem. Eur.
J. 2001 7, 1266.
(5) Lambert, C.; No¨ll, G. Angew. Chem., Int. Ed. 1998, 37, 2107. (b) Lambert,
C.; No¨ll, G. J. Am. Chem. Soc. 1999, 121, 8434. (c) Lambert, C.; No¨ll, G.;
Schelter, J. Nat. Mater. 2002, 1, 69.
(6) (a) Coropceanu, V.; Malagoli, M.; Andre´, J. M.; Bre´das, J. L. J. Chem.
Phys. 2001, 115, 10409. (b) J. Am. Chem. Soc. 2002, 124, 10519. (c) Theor.
Chem. Acc. 2003, 110, 59.
(7) Coropceanu, V.; Lambert, C.; No¨ll, G.; Bre´das, J. L. Chem. Phys. Lett.
2003, 373, 153.
(8) Risko, C.; Barlow, S.; Coropceanu, V.; Halik, M.; Bre´das, J. L.; Marder,
S. R. Chem. Commun. 2003, 194.
µabEop
Hab
)
(1)
eR
Here, R is the effective separation distance between the donor
and acceptor sites (diabatic states).12
(9) Kido, J.; Kimura, M.; Nagai, K. Science 1995, 267, 1332.
(10) (a) Launay J. P. Chem. Soc. ReV. 2001, 30, 386. (b) Demadis, K. D.;
Hartshorn, C. M.; Meyer, T. J. Chem. ReV. 2001, 101 2655. (c) Brunschwig,
B. S.; Creutz, C.; Sutin, N. Chem. Soc. ReV. 2002, 31 168.
9
10.1021/ja039263u CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 2727-2731
2727

A R T I C L E S
Coropceanu et al.
Figure 1. Chemical structures of 1,4-bis(dimethylamino)benzene (1); 4,4′-
bis(dimethylamino)biphenyl (2); 1,4-bis(diphenylamino)benzene (3); 1,4-
bis[di(4-methoxyphenyl)amino]benzene (4); 4,4′-bis(diphenylamino)biphe-
nyl (5); and 4,4′-bis[di(4-methoxyphenyl)amino]biphenyl (6).
Figure 2. Sketch of typical adiabatic energy surfaces for (a, left) a class-
II system in which the horizontal axis represents the antisymmetric
vibrational mode Q- (in a class-III system, all three surfaces exhibit only
one minimum at Q- ) 0) and (b, right) a class-II or -III system as a function
of the symmetric vibrational coordinate Q+.
Another widely applied experimental method to probe the
electronic coupling in organic MV systems is based on the rate
of electron exchange between redox centers measured by
electron spin resonance (ESR) spectroscopy.1-3 It has been
recently shown that in bis(hydrazine) radical cations3b the Hab
values obtained from optical data accurately predict the in-
tramolecular electron-transfer rate constants determined by ESR.
However, in several other MV systems, such as those including
perchlorinated triphenylmethyl redox sites, interpretation of ESR
results requires markedly larger values for Hab than those
obtained from optical measurements.1c While the discrepancy
between the results might be simply due to the use of erroneous
values for R, it has also been suggested that the use of Hush
theory (eq 1) to calculate Hab may not be correct for strongly
coupled organic compounds.1c
Many organic MV systems, such as triarylamine-based
compounds, have been found to display significant electronic
couplings and are intermediate between class II and class III or
even belong to class III in the classification scheme of Robin
and Day.13 Such strong electronic interactions generally give
rise to electron-transfer rates much faster than the typical time
scales of magnetic resonance; as a result, the electronic coupling
in these systems can often be obtained only from IV-CT band
measurements. Thus, in such instances, it is important to obtain
an alternative experimental estimate of the electronic couplings.
This would not only provide a better understanding of the
electron-transfer phenomena but will help to establish the
reliability of the results derived from IV-CT bands.
systems, and 3+-6+ are class-II/class-III borderline systems (see
Figure 2). The UPS results are compared with those derived
previously from optical data and from electronic structure
calculations.
Experimental and Computational Section
Samples of 1 (98% purity) and 2 (97% purity) were purchased from
Aldrich. Compounds 3-6 were synthesized according to published
procedures.5,14 Gas-phase photoelectron spectra of 1-6 were collected
in the Center for Gas-Phase Electron Spectroscopy (Department of
Chemistry, The University of Arizona), using an instrument and
experimental procedures described in more detail elsewhere.15a The
samples sublimed (10-4 Torr) at 28-36 (1), 130-146 (2) 180-200
(3), 230-270 (4), 220-250 (5), and 260-290 °C (6) with no evidence
of contaminants present in the gas phase during data collection.
Instrument resolution during data collection was better than 30 meV
(measured using the full width at half-height for the 2P3/2 ionization of
Ar). Figures 3 and 4 show the experimental data; in these figures the
vertical length of each data mark represents the experimental variance
of that point.15b The valence ionization bands are represented analytically
with the best fit of asymmetric Gaussian peaks. Each Gaussian peak is
defined by four degrees of freedom: position, intensity, half-width to
the high-binding-energy side, and half-width to the low-binding-energy
side. The widths listed in Table 1 are the sum of the two individual
half-widths. The peak positions and half-widths are reproducible to
about (0.02 eV.
In this contribution, we report the results of direct measure-
ments of the electronic coupling parameters in amine-based MV
systems (see Figure 1) using gas-phase ultraviolet photoelectron
spectroscopy (UPS); 1+ and 2+ are examples of class-III
Solutions containing 5+ were generated by the addition of excess
solid 5 to solutions of [(p-BrC6H4)3N]+[SbCl6]- (Aldrich). The IV-CT
maximum was observed at 1395 nm (7166 cm-1; ꢀmax ) 27 900 M-1
cm-1) in dry dichloromethane, with an additional maximum at 484 nm
(20 667 cm-1; ꢀmax ) 20 900 M-1 cm-1). In acetonitrile, the maxima
(11) (a) Hush, N. S. Prog. Inorg. Chem. 1967, 8, 391. (b) Hush, N. S. Coord.
Chem. ReV. 1985, 64, 135. (c) Creutz, C.; Newton, M. D.; Sutin, N. J.
Photochem. Photobiol., A 1994, 82, 47. (d) Cave, R. J.; Newton, M. D.
Chem. Phys. Lett. 1996, 249, 15.
(12) While the parameters µab and Eop are accessible from optical measurements,
the diabatic transfer distance R cannot in principle be directly measured.
Only significantly after the seminal work of Hush, Cave, and Newton11d
was a prescription developed on how to obtain R. In addition to µab, this
model requires knowledge of the adiabatic dipole moment shift that can
be obtained, in principle, from Stark effect experiments or from quantum-
mechanical calculations.
(13) (a) Robin and Day classified the MV systems into three types: class I,
completely valence-trapped (no electronic coupling between redox sites);
class II, valence-trapped (weak coupling); and class III, delocalized valency
(strong coupling).13b (b) Robin, M. B.; Day, P. AdV. Inorg. Chem.
Radiochem. 1967, 10, 247.
(14) The known compound 5 (Madelung, R. Justus Liebigs Ann. Chem. 1927,
454, 36) was prepared from diphenylamine and 4,4′-dibromobiphenyl using
standard palladium-catalyzed methodology (Thayumanavan, S.; Barlow,
S.; Marder, S. R. Chem. Mater. 1997, 9, 3231), and its identity and purity
were checked by NMR and mass spectroscopy (1H NMR (CD2Cl2, 300
MHz) δ 7.47 (d, 4H, J ) 8.5 Hz), 7.27 (t, 8H, J ) 8.4 Hz), 7.10 (m, 12H),
7.03 (t, 4H, J ) 7.3 Hz); 13C NMR (CD2Cl2, 75 MHz) δ 148.1, 147.2,
135.0, 129.6, 127.6, 124.7, 124.4, 123.2; FAB HRMS calcd for C36H28N2:
488.2252; found: 488.2243).
(15) (a) Cornil, J.; Gruhn, N. E.; dos Santos, D. A.; Malagoli, M.; Lee, P. A.;
Barlow, S.; Thayumanavan, S.; Marder, S. R.; Armstrong, N. R.; Bre´das,
J. L. J. Phys. Chem. A 2001, 105, 5206. (b) Lichtenberger, D. L.;
Copenhaver, A. S. J. Electron Spectrosc. Relat. Phenom. 1990, 50, 335.
9
2728 J. AM. CHEM. SOC. VOL. 126, NO. 9, 2004

Couplings in Organic Mixed-Valence Compounds
A R T I C L E S
Table 1. Fit Parameters from a Deconvolution with Gaussians of
the First Two UPS Ionizations and Experimental and KT-AM1 as
Well as TD-DFT Theoretical Values (Obtained at the
DFT-Optimized Geometries of the Neutral State) of the Parameter
∆IP (All Values in electronvolts)
IP position (width) IP position (width) ∆IP(UPS) ∆IP(KT-AM1) ∆IP(TD-DFT)
1
2
1
2
3
4
5
6
6.78 (0.74)
8.36 (0.34)
1.58
0.83
0.65
0.69
0.30
0.38
1.83
0.84
0.63
0.59
0.37
0.35
1.99
1.25
1.05
0.99
0.76
0.73
6.75 (0.65)
6.70 (0.60)
6.36 (0.59)
6.79 (0.50)
6.50 (0.64)
7.58 (0.54)
7.35 (0.36)
7.05 (0.49)
7.09 (0.30)
6.88 (0.62)
Hab were calculated by means of time-dependent DFT (TD-DFT)18 at
the same B3LYP level. In addition, Hab was derived from Koopmans’
theorem19 using Kohn-Sham-type molecular orbitals and Hartree-
Fock-type orbitals obtained at both ab initio and semiempirical Austin
Model 1 (AM1) levels. All nonempirical calculations were carried out
with the 6-31G** split valence plus polarization basis set20 using the
Gaussian 98 suite of programs.21
Results and Discussion
In a Ma-BR-Mb system, the splitting between the first and
second ionization potentials (IP), ∆IP ) IP2 - IP1, is related to
the strength of the electronic coupling in the MV Ma+-BR-
Mb state.11b,22,23 Indeed, since ∆IP is determined at the geometry
of the neutral molecule, and consequently at a symmetric radical-
cation geometry configuration, the first and second IPs cor-
respond to creation of the delocalized electronic states:
Figure 3. Gas-phase photoelectron spectra of systems 1-6.
Ψ+ ) (1/x2)(Ψa + Ψb)
Ψ- ) (1/x2)(Ψa - Ψb)
(2)
of the radical cation. Bearing in mind that, in the framework of
the two-state model (see Figure 2a), the energy difference,
∆0 ) ꢀ - ꢀ , between the molecular electronic states Ψ+ and
+
-
Ψ-, is associated with 2Hab, then ∆IP gives a direct estimate
of the electronic coupling integral, Hab) ∆IP/2.
In triarylamine MV systems such as 3+-6+, it has been
shown that the electron-transfer (ET) process is significantly
coupled not only to antisymmetric vibrations Q- (which are
the only vibrations that enter as reaction coordinate in the
conventional Hush-Marcus model)11,24 but also to the totally
(18) (a) Casida, M. E.; Jamorski, C.; Casida, K. C.; Salahub, D. R. J. Chem.
Phys. 1998, 108, 4439. (b) Stratmann, R. E.; Scuseria, G. E.; Frisch, M. J.
J. Chem. Phys. 1998, 109, 8218.
(19) Koopmans, T. Physica 1933, 1, 104.
(20) (a) Ditchfield, R.; Hehre, W. J.; Pople, J. A. J. Chem. Phys. 1971, 54, 724.
(b) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972, 56,
2257. (c) Hariharan, P. C.; Pople, J. A. Mol. Phys. 1974, 27, 209. (d)
Gordon, M. S. Chem. Phys. Lett. 1980, 76, 163. (e) Hariharan, P. C.; Pople,
J. A. Theor. Chim. Acta. 1973, 28, 213.
Figure 4. High-resolution close-up of the first ionizations of the benzene-
bridged systems 1, 3, and 4, and the biphenyl-bridged systems 2, 5, and 6.
were observed at 1279 nm (7816 cm-1) and 479 nm (20861 cm-1), but
the cation was too unstable to allow accurate determination of ꢀmax
values. The IV-CT data of 1+-4,+ and 6+ have been reported
elsewhere.3,5-7
As in the case of our previous calculations on 3, 4, and 6,6,7 the
geometry of molecules 1, 2, and 5 were optimized at the density
functional theory (DFT) level, using the B3LYP functionals, where
Becke’s three-parameter hybrid exchange functional is combined with
the Lee-Yang-Parr correlation functional.16,17 The electronic couplings
(21) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann,
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,
K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,
J. V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.;
Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham,
M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-
Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.9;
Gaussian, Inc.: Pittsburgh, PA, 1998.
(22) Paddon-Row, M. N. Acc. Chem. Res. 1982, 15, 245.
(16) (a) Becke, A. D. Phys. ReV. A 1988, 38, 3098. (b) Becke, A. D. J. Chem.
Phys. 1993, 98, 5648.
(17) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785.
(23) Jordan, K. D.; Paddon-Row, M. N. Chem. ReV. 1992, 92, 395.
(24) (a) Marcus, R. A. Discuss. Faraday Soc. 1960, 29, 21. (b) Marcus, R. A.;
Sutin, N. Comments Inorg. Chem. 1986, 5, 119.
9
J. AM. CHEM. SOC. VOL. 126, NO. 9, 2004 2729

A R T I C L E S
Coropceanu et al.
Table 2. Electronic Coupling Hab (eV) for Systems 1+-6+ As
symmetric modes Q+.6,7,25,26 As a consequence, the effective
electronic coupling depends on the coordinate of the symmetric
vibration(s)7 (see Figure 2b); therefore, using the ∆IP value will
give a lower limit estimate for Hab.
Obtained from Optical and UPS Measurements
optical
∆IP/2
(∆IP/2) + La
1+
2+
3+
4+
5+
6+
1.01b,c
0.79
0.42
0.33
0.35
0.15
0.19
-
-
0.61b,c
The gas-phase photoelectron spectra of 1-6 are shown in
Figure 3; the details of the first ionizations are shown in Figure
4. From these figures and Table 1, it is seen that replacing the
methyl groups of 1 and 2 with phenyl groups in 3 and 5 has
little effect on the first ionizations IP1 but leads to very
significant reductions in ∆IP splitting; on the other hand,
methoxy substitution on the terminal rings when going to 4 and
6 causes a 0.3 eV decrease in IP1 but has little effect on ∆IP
with respect to 3 and 5. The replacement of the benzene bridge
with a biphenyl bridge of approximately twice the length reduces
the ∆IP splitting, and thus the electronic coupling, in the same
proportion, i.e., by about 50%.
0.70 (0.45)d,e
0.53 (0.40)d,f
0.36 (0.20)d,g
0.31 (0.19)d,h
0.66
0.55
0.30
0.34
a Using the following values of L (obtained either from electronic
structure calculations or the fit of the profile of IV-CT bands): 0.3 eV
(3+),7 0.2 eV (4+),6b 0.15 eV (5+), and 0.15 eV (6+).6b,7 b Calculated taking
Hab ) Eop/2. c From ref 3f; in CH3CN. d Calculated according to eq 1 using
R ) R(TD-DFT) and, between parentheses, R ) R(NN); both R(TD-DFT)
and R(NN) values are obtained using the DFT-optimized geometries of the
radical-cation state.6,7 e From ref 7: Eop ) 1.50 eV; µab ) 8.1 D in
propionitrile/0.2 M (TBA)H. f From ref 5b: Eop ) 1.18 eV; µab ) 9.2 D in
CH2Cl2. g Present work: Eop ) 0.89 eV; µab ) 11.0 D in CH2Cl2. h From
ref 5b: Eop ) 0.79 eV; µab ) 11.6 D in CH2Cl2.
We have recently evaluated ∆0 (and thus ∆IP) using several
computational approaches: (a) from the energy spectrum
obtained at the TD-DFT level, (b) from DFT calculations by
computing the energy of the lowest state of each symmetry
representation, and (c) by means of Koopmans’ theorem (KT).6
It was found that the best agreement with experimental data
was obtained from KT-AM1 calculations; the KT-AM1 values
are, therefore, given in Table 1. In the context of electronic
structure calculations, it is also particularly interesting to
examine the reliability of the TD-DFT results, since this method
is currently widely used for excited-state computations. In
contrast to the simple semiempirical AM1 calculations, B3LYP-
TD-DFT overestimates the ∆0 value for all systems; for instance
in 5+ and 6+, the TD-DFT estimates are about 100% larger
than the experimental value. This result is in line with current
findings that, due to inherent problems in the applied standard
exchange-correlation functionals (see for instance ref 27 for
detailed explanations), TD-DFT yields substantial errors for
systems exhibiting charge-transfer states.27 We note, however,
that for all systems studied here, the error of the TD-DFT
estimate is systematic, being about 0.4 eV; this finding suggests
that it should be possible to obtain more accurate TD-DFT
estimates via an empirical correction.28
significantly lower than the estimates based on the direct
geometric distances between redox centers. For instance, it was
shown very recently3g that in delocalized nitro-centered MV
systems, R is only about 25-40% of the direct distance between
the nitrogen atoms, R(NN). Since for 3+-6+ there are no
experimental electron-transfer distances available, Table 2 gives
values of Hab obtained using the following: (i) TD-DFT
estimates for the ET distances R ) R(TD-DFT)6,7 and (ii) the
commonly applied approach, where R is approximated to the
nitrogen-nitrogen distance, R ) R(NN). In 3+-6+, the ratio
R(TD-DFT)/R(NN) is about 0.6;6,7 we expect, therefore, the Hab
estimates based on the R(TD-DFT) values to be more realistic.
The data in Table 2 indicate that the Hab values estimated
optically show trends similar to the ∆IP/2 values and are
comparable in magnitude. However, as expected from the
neglect of the interaction with symmetric vibrations (see Figure
2b), the UPS data yield values lower than those derived from
optical data. The strength of the electron-vibration couplings
with symmetric modes can actually be estimated from the data
for 1+ and 2+: since 1+ and 2+ are class-III systems, the
geometries of the ground states of both cation and neutral species
possess a symmetric configuration; the difference32 of 0.38-
0.44 eV between Eop and ∆IP can then be related to the
relaxation processes along symmetric vibrations.
We now turn to a comparison of the electronic couplings as
derived from UPS measurements and obtained from optical
data;29 see Table 2. Hab for 1+ and 2+ was determined as Hab
)
To obtain more accurate estimates of Hab from UPS data,
the ∆IP/2 values should be corrected to account for the
interaction with symmetric vibrations. The most accurate way
to accomplish such a correction would be to fit the shape of
the ionization spectrum to a vibronic coupling model. However,
a simple but effective estimate of the vibronic contribution can
Eop/2,30 while for systems 3+-6+ Hab was derived using eq 1.
In the latter cases, the question of the appropriate value of R
arises; several experimental and theoretical studies3g,6,7,31 indicate
that the actual electron-transfer distances in organic systems are
(25) Similar results have been recently found for other delocalized organic MV
systems (see refs 3a,g and 8).
be provided by correcting the UPS values according to Hab
)
(∆IP/2) + L, where L is the vibrational relaxation energy of
the symmetric modes associated with the excitation of the cation
from its ground state to its first excited state (see Figure 2b).
This approximation assumes that the minimum of the cation
(26) (a) Interestingly, the fact that the interaction with symmetric vibrations
could play a major role in the charger-transfer process in strongly
delocalized MV systems was also predicted by Hush.26b,c (b) Hush, N. S.
In Mixed-Valence Compounds: Theory and Applications in Chemistry,
Physics, Geology and Biology; Brown, D. B., Ed.; D. Reidel Academic
Publishers: Dordrecht, The Netherlands, 1980. (c) Reimers, R.; Hush, N.
S. Chem. Phys. 1996, 208, 177.
(27) (a) Drew, A.; Weisman, J. L.; Head-Gordon, M. J. Chem. Phys. 2003, 119,
2943. (b) Grimme, S.; Parac, M. ChemPhysChem 2003, 3, 292.
(28) (a) Recent28b investigations of (neutral) heterocyclic conjugated polymers
show that after an empirical correction via a linear regression, TD-DFT
yields accurate results for band gap. (b) Hutchison, G. R.; Ratner, M. A.;
Marks, T. J., J. Phys. Chem. A 2002, 106, 1059.
(29) It should be noted that the optical values apply to solution, whereas the
UPS values are determined in the gas phase. However, in strongly coupled
systems, the IV-CT bands show typically little dependence on solvent
polarity; this suggests that the medium has little effect on Hab and, therefore,
that solution and gas-phase values should be similar.
(30) (a) In the case of strongly coupled MV systems (class III), the energy of
the intervalence transition becomes a direct measure of the electronic
coupling: Eop ) 2Hab.10c The observed vibrational structure in the IV-CT
band of 1+ 3f and 2+ 30b and the absence of solvent dependence3f are features
characteristic of a class-III system. (b) Shida, T. Electronic Absorption
Spectra of Radical Ions; Elsevier: Amsterdam, 1988.
(31) (a) Nelsen, S. F.; Newton, M. D. J. Phys. Chem. A 2000, 104, 10023. (b)
Johnson, R. C.; Hupp, J. T. J. Am. Chem. Soc. 2001, 123, 2053.
(32) A similar difference between Eop and ∆IP values has been recently reported
for a delocalized phthalocyanine MV dimer: Binstead, R. A.; Reimers, J.
R.; Hush, N. S. Chem. Phys. Lett. 2003, 378, 654.
9
2730 J. AM. CHEM. SOC. VOL. 126, NO. 9, 2004

Couplings in Organic Mixed-Valence Compounds
A R T I C L E S
excited-state surface corresponds to a geometry similar to that
of the neutral ground state and that the two cation potential
surfaces have similar curvatures (this is justified by our previous
work on triarylamine systems which suggests that the geometry
of the cation in its first excited state is indeed similar to that of
the neutral state).6 The new estimates of Hab derived in this
manner are now in very good agreement with those calculated
by a Hush analysis of the IV-CT bands using R(TD-DFT); as
expected, they are larger than the Hush values obtained by
setting R to the nitrogen-nitrogen distance. The derived values
(see Table 2) of 0.66, 0.55, and 0.32 eV for 3+, 4+, and 6+
obtained using this approximation compare well with the
estimates of 0.69, 0.60, and 0.38 eV obtained previously for
these systems from the vibronic coupling simulations of the IV-
CT bands.6,7
Assuming that 3+-6+ are class-II systems, we have evaluated
the activation barrier, ∆Gq for the thermal electron transfer
between the two redox centers of the triarylamine systems
according to ∆Gq ) (Eop - 2HAB)2/4Eop. Combining the IV-
CT values for Eop with the UPS-corrected values of Hab, we
obtain for all systems rather small barrier estimates. For
example, ∆Gq is ca. 7 meV for 6+, indicating that these species
are close to the class-II/class-III borderline. However, due to
the remaining uncertainties in estimating Hab, the possibility that
3+ and 4+ are in fact class-III systems cannot be ruled out. The
latter possibility is also supported by the vibronic coupling
simulation of the IV-CT band;6,7 in addition, very recent X-ray
crystal structure studies of 3+ and a derivative of 5+ reveal that
the ground state in both MV systems are characterized by a
symmetric geometry.33,34
Conclusions
In this work, we have demonstrated that the electronic
coupling integrals in strongly coupled organic MV systems can
be efficiently probed by means of gas-phase photoelectron
spectroscopy. We expect, therefore, that this technique will
receive more frequent applications in the investigations of ET
processes.
Our results provide an important verification that the Hush
model can be used to describe organic MV systems close to
the class-II/class-III borderline. The present study also confirms
the significant coupling of ET processes with symmetric modes
in triarylamine-based mixed-valence systems.
Acknowledgment. The work at the Georgia Institute of
Technology is supported by the National Science Foundation,
through the STC Program under Award No. DMR-0120967 and
through Grant CHE-0342321, and the IBM Shared University
Research Program. The work in Wu¨rzburg is supported by the
Deutsche Forschungsgemeinschaft and the Fonds der Chemis-
chen Industrie as well as by JASCO GmbH Deutschland. The
authors also thank S. Thayumanavan for preparing the sample
of 5.
(33) Amthor, S.; Engel, V.; Erdmann, M.; Kriegisch, V.; Lambert, C.; Leusser,
D.; No¨ll, G.; Popp, J.; Schelter, J.; Schmitt, M.; Stahl, R.; Stalke, D.;
Szeghalmi, A.; Zabel, M. Manuscript in preparation.
(34) Low, P. J.; Paterson, M. A. J.; Puschmann, H.; Goeta, A. E.; Howard, J.
A. K.; Lambert, C.; Cherryman, J.; Tackley, D. R.; Leeming, S.; Brown,
B. Chem. Eur. J. 2004, 10, 83.
JA039263U
9
J. AM. CHEM. SOC. VOL. 126, NO. 9, 2004 2731