Electron affinities of 1,1-diaryl-2,3,4,5-tetraphenylsiloles: Direct measurements and comparison with experimental and theoretical estimates

Article abstract of DOI:10.1021/ja051139i

We present a comprehensive experimental and theoretical characterization of the electronic structure of four 1,1-diaryl-2,3,4,5-tetraphenylsiloles (aryl = phenyl, 2-(9,9-dimethylfluorenyl), 2-thienyl, pentafluorophenyl). Solid-state electron affinities and ionization potentials of these siloles were measured using inverse-photoelectron spectroscopy (IPES) and photoelectron spectroscopy (PES), respectively; the density of electronic states obtained from calculations performed at the density functional theory (DFT) level corresponds very well to the PES and IPES data. The direct IPES measurements of electron affinity were then used to assess alternative estimates based on electrochemical and/or optical data. We also used DFT to calculate the reorganization energies for the electron-transfer reactions between these siloles and their radical anions. Additionally, optical data and ionization potential and electron affinity data were utilized to estimate the binding energies of excitons in these siloles.

Full text of DOI:10.1021/ja051139i

Published on Web 06/04/2005
Electron Affinities of 1,1-Diaryl-2,3,4,5-tetraphenylsiloles:
Direct Measurements and Comparison with Experimental and
Theoretical Estimates
Xiaowei Zhan,† Chad Risko,† Fabrice Amy,‡ Calvin Chan,‡ Wei Zhao,‡
Stephen Barlow,† Antoine Kahn,*,‡ Jean-Luc Bre´das,*,† and Seth R. Marder*,†
Contribution from the School of Chemistry and Biochemistry and the Center for Organic
Photonics and Electronics, Georgia Institute of Technology, Atlanta, Georgia 30332, and
Department of Electrical Engineering, Princeton UniVersity, Princeton, New Jersey 08544
Received February 22, 2005; E-mail: seth.marder@chemistry.gatech.edu
Abstract: We present a comprehensive experimental and theoretical characterization of the electronic
structure of four 1,1-diaryl-2,3,4,5-tetraphenylsiloles (aryl ) phenyl, 2-(9,9-dimethylfluorenyl), 2-thienyl,
pentafluorophenyl). Solid-state electron affinities and ionization potentials of these siloles were measured
using inverse-photoelectron spectroscopy (IPES) and photoelectron spectroscopy (PES), respectively; the
density of electronic states obtained from calculations performed at the density functional theory (DFT)
level corresponds very well to the PES and IPES data. The direct IPES measurements of electron affinity
were then used to assess alternative estimates based on electrochemical and/or optical data. We also
used DFT to calculate the reorganization energies for the electron-transfer reactions between these siloles
and their radical anions. Additionally, optical data and ionization potential and electron affinity data were
utilized to estimate the binding energies of excitons in these siloles.
I. Introduction
with the hole-transport material N,N′-diphenyl-N,N′-bis(1-naph-
thalenyl)-(1,1′-biphenyl)-4,4′-diamine (NPD) and has been
exploited in a three-layer device with the configuration Ag:
Mg/1/4/NPD/ITO, which has an external EL quantum efficiency
of 3.4%.6
Siloles (silacyclopentadienes, Figure 1) have recently attracted
attention as useful materials for organic electronics since they
may exhibit high electron mobilities and high photolumines-
cence quantum yields. Time-of-flight electron mobilities of 1
were found to be ca. 2 × 10-4 cm2/Vs, a 2-order of magnitude
improvement compared to the well-established electron-transport
(ET) material tris(8-hydroquinolinato)aluminum (Alq3); more-
over, the electron transport was shown to be nondispersive and
independent of exposure to air.1-3 Accordingly, siloles have
been used as efficient ET materials in organic light-emitting
diodes (OLEDs).4 Furthermore, 2 possesses a photoluminescence
(PL) quantum yield of 97 ( 3% in the solid state and has been
used as an effective lumophore in OLEDs; for example, an
OLED using 2 as an emissive layer and 1 as an ET layer shows
a very low operating voltage of 2.5 V and a high external
electroluminescence (EL) quantum efficiency of 4.8%.1 Another
EL device based on 3 shows an excellent external EL quantum
efficiency of 8%,5 while 4 has been shown to form a highly
emissive exciplex (PL quantum yield of 62%) at the interface
In designing organic electronic devices, it is important to be
able to assess hole- and electron-injection barriers at the various
organic/organic and organic/inorganic interfaces. These are
typically estimated by comparing ionization potentials (IPs) and
electron affinities (EAs) of the molecular components with each
other and/or with the Fermi energies of inorganic electrode
materials, respectively.7 Although approximate IPs and EAs
have often been derived from solution electrochemical data, such
evaluations require reversible electrochemistry and assume
similar solvation effects and solid-state polarization effects on
the formation of ions in both the species of interest and the
reference compound. The most direct experimental probes of
IP and EA are photoelectron spectroscopy (PES) and inverse-
photoelectron spectroscopy (IPES), respectively.
Here we report a study of the electronic properties of 1,1-
diaryl-2,3,4,5-tetraphenylsiloles. We selected this class of siloles
given that: (i) it has previously been reported that OLEDs based
on siloles with R1 ) R1′ ) Ph are much brighter and more
† Georgia Institute of Technology.
‡ Princeton University.
(1) Murata, H.; Kafafi, Z. H.; Uchida, M. Appl. Phys. Lett. 2002, 80, 189-
191.
(6) Palilis, L. C.; Ma¨kinen, A. J.; Uchida, M.; Kafafi, Z. H. Appl. Phys. Lett.
2003, 82, 2209-2211.
(2) Murata, H.; Malliaras, G. G.; Uchida, M.; Shen, Y.; Kafafi, Z. H. Chem.
Phys. Lett. 2001, 339, 161-166.
(7) This approach overlooks effects such as vacuum-level offsets brought about
due to interface dipole formation and effects due to creation of new species
with “gap states” close to the interfaces. Indeed, Kafafi and co-workers
have found gap states formed at the silole/magnesium interface using PES.
However, these effects are likely to be negligible in organic/organic
interfaces and will be very material-specific in the case of organic/inorganic
interfaces; see ref 3.
(3) Ma¨kinen, A. J.; Uchida, M.; Kafafi, Z. H. J. Appl. Phys. 2004, 95, 2832-
2838.
(4) Tamao, K.; Uchida, M.; Izumizawa, T.; Furukawa, K.; Yamaguchi, S. J.
Am. Chem. Soc. 1996, 118, 11974-11975.
(5) Chen, H. Y.; Lam, W. Y.; Luo, J. D.; Ho, Y. L.; Tang, B. Z.; Zhu, D. B.;
Wong, M.; Kwok, H. S. Appl. Phys. Lett. 2002, 81, 574-576.
9
10.1021/ja051139i CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 9021-9029
9021

A R T I C L E S
Zhan et al.
rophenyl (V) and compared them to the previously reported 1,1-
diphenyl-2,3,4,5-tetraphenylsilole, II, and 1,1-methyl-2,3,4,5-
tetraphenylsilole, I. As part of this study, we report the first
EA measurements for solid-state films of siloles. In addition,
we report solid-state IP data, electrochemical data, UV-vis
absorption and emission data, and the results of quantum-
chemical calculations.
II. Experimental and Theoretical Methodolgy
Synthesis and Characterization. The 1H and 13C NMR spectra were
measured on a Varian Mercury 300 spectrometer using tetramethylsilane
(TMS; δ ) 0 ppm) as an internal standard. Mass spectra were measured
on a VG Instruments 70-SE using the electron impact (EI) mode.
Elemental analyses were carried out by Atlantic Microlabs using a
LECO 932 CHNS elemental analyzer; fluorine analysis was performed
by flask combustion followed by ion chromatography. Solution
(chloroform, hexane, and toluene) and thin-film (on quartz substrate)
UV-vis absorption spectra were recorded on a Varian Cary 5E UV-
vis-NIR spectrophotometer, while solution (chloroform) and thin-film
PL spectra were recorded on a Shimadzu FP-5301PC spectrofluorom-
eter. Electrochemical measurements were carried out under nitrogen
on a deoxygenated solution of tetra-n-butylammonium hexafluorophos-
phate (0.1 M) in 1:1 (by volume) acetonitrile/toluene using a computer-
controlled BAS 100B electrochemical analyzer, a glassy-carbon working
electrode, a platinum-wire auxiliary electrode, and a Ag wire anodized
with AgCl as a pseudo-reference electrode. Potentials were referenced
to the ferrocenium/ferrocene (FeCp2+/0) couple by using ferrocene as
an internal standard.
Diphenylacetylene, silicon tetrachloride, lithium wire, n-butyllithium
(1.6 M in hexane), 2-bromofluorene, iodomethane, 2-bromothiophene,
bromopentafluorobenzene, and magnesium shavings were purchased
from Aldrich and used as received without further purification. 1,1-
Dimethyl-2,3,4,5-tetraphenylsilole (I) and 1,1-diphenyl-2,3,4,5-tet-
raphenylsilole (II) were prepared according to published procedures.8
1,1-Dichloro-2,3,4,5-tetraphenylsilole (VI). This intermediate was
prepared according to a procedure reported by Chen et al.14 Dipheny-
lacetylene (16.8 mmol, 3.0 g) and clean lithium shavings (14 mmol,
98 mg) were put in a 100-mL three-necked round-bottom flask and
deoxygenated with nitrogen for 30 min. Dry THF (15 mL) was then
added. The reaction mixture was stirred at room temperature for 14 h
under nitrogen atmosphere. The resultant deep green mixture was
diluted with 50 mL of dry THF and added dropwise to a solution of
silicon tetrachloride (6.3 mmol, 1.07 g) in 20 mL of dry THF over a
period of 0.5 h at room temperature. The brown mixture was stirred at
the same temperature for 2 h and then refluxed for 5 h to give a dark
yellow solution. The THF solution of the resultant reactive intermediate
VI was used in situ for the synthesis of siloles III-V without isolation.
2-Bromo-9,9-dimethylfluorene (VII). 2-Bromofluorene (160.0
mmol, 39.2 g) was dissolved in 300 mL of DMSO at 60 °C. Potassium
iodide (16.0 mmol, 2.7 g) and iodomethane (360.0 mmol, 51.1 g) were
then added; finally, powdered potassium hydroxide (640.0 mmol, 36.0
g) was slowly added. The mixture was stirred at room temperature
overnight. The mixture was dropped into 2000 mL of water, and a
large amount of light yellow precipitate formed. The solid was filtered,
washed with water, and dried under vacuum. The light yellow solid
was purified by recrystallization from methanol to give white crystals
(40.3 g, 92%). mp: 56 °C. 1H NMR (300 MHz, CDCl3): δ 7.72-7.65
(m, 1H), 7.60-7.54 (m, 2H), 7.48-7.39 (m, 2H), 7.37-7.30 (m, 2H),
1.47 (s, 6H). 13C NMR (75 MHz, CDCl3): δ 155.63, 153.19, 138.18,
138.11, 130.04, 127.63, 127.14, 126.12, 122.62, 121.36, 120.98, 120.03,
47.07, 26.98. HRMS (EI): m/z 272.0197 (calcd for C15H13Br, 272.0201).
Anal. Calcd for C15H13Br: C, 65.95; H, 4.80. Found: C, 65.96; H,
4.80.
Figure 1. Structures of siloles discussed in this article. 1-6 are discussed
in the Introduction; I-V are the species compared in detail in this article.
efficient than those based upon siloles with R1 ) R1′ ) alkyl8
and (ii) thermal stability is an important materials parameter
for OLEDs, and 1,1-diaryl siloles have higher melting points
and glass-transition temperatures than their 1,1-dialkyl coun-
terparts (see Table S1 in Supporting Information). The electronic
effects of 1,1-substituents in siloles have previously been probed
by electrochemistry,9,10 as well as by UV-vis absorption and
ab initio calculations;11 however, there is only one report in
which two different aryl groups are compared (5 versus 6), and
this study only compared UV-vis data.11 Additionally, despite
considerable interest in the ET properties of siloles, no direct
measurements of EAs have been published, and only a few
studies have involved the determination of an IP (1, 2, 4) by
PES.3,12,13 To probe the effects of the 1,1-diaryl groups on the
electronic structure of the siloles in more detail, we have
synthesized three new species where these aryl groups are
2-(9,9-dimethylfluorenyl) (III), 2-thienyl (IV), or pentafluo-
(8) Tang, B. Z.; Zhan, X. W.; Yu, G.; Lee, P. P. S.; Liu, Y. Q.; Zhu, D. B. J.
Mater. Chem. 2001, 11, 2974-2978.
(9) Dhiman, A.; Zhang, Z. R.; West, R.; Becker, J. Y. J. Electroanal. Chem.
2004, 569, 15-22.
(10) Ferman, J.; Kakareka, J. P.; Klooster, W. T.; Mullin, J. L.; Quattrucci, J.;
Ricci, J. S.; Tracy, J. J.; Vining, W. J.; Wallace, S. Inorg. Chem. 1999, 38,
2464-2472.
(11) Yamaguchi, S.; Jin, R. Z.; Tamao, K. J. Organomet. Chem. 1998, 559,
73-80.
(12) Ma¨kinen, A. J.; Uchida, M.; Kafafi, Z. H. Appl. Phys. Lett. 2003, 82, 3889-
3891.
(13) Palilis, L. C.; Ma¨kinen, A. J.; Murata, H.; Uchida, M.; Kafafi, Z. H. Proc.
SPIE 2003, 4800, 256-270.
(14) Chen, J.; Law, C. C. W.; Lam, J. W. Y.; Dong, Y.; Lo, S. M. F.; Williams,
I. D.; Zhu, D.; Tang, B. Z. Chem. Mater. 2003, 15, 1535-1546.
9
9022 J. AM. CHEM. SOC. VOL. 127, NO. 25, 2005

EAs of 1,1-Diaryl-2,3,4,5-tetraphenylsiloles
A R T I C L E S
1,1-Bis(9,9-dimethylfluoren-2-yl)-2,3,4,5-tetraphenylsilole (III). To
a 250-mL three-necked round-bottom flask were added VII (9.5 mmol,
2.6 g) and 20 mL of dry diethyl ether; the solution was deoxygenated
with nitrogen for 30 min and cooled to -78 °C. n-BuLi (9.6 mmol, 6
mL, 1.6 M) in hexane was dropwise added over a period of 10 min.
The mixture was stirred at the same temperature for 1 h to give a white
suspension. The THF solution of VI was cooled to 0 °C and transferred
portionwise into the above organolithium suspension over a period of
30 min. The brown mixture was stirred for 1 h at -78 °C and was
gradually warmed to room temperature; the solution was stirred
overnight. The resulting yellow-green solution was washed with water;
the organic layer was extracted with ether and dried over anhydrous
MgSO4. The solvent was removed, and the residue was purified by
flash column chromatography over silica gel using hexane/dichlo-
romethane ) 8/1 as eluent. Recrystallization from ethanol gave faintly
greenish-yellow and highly blue fluorescent crystals (1.6 g, 44%, based
on 2-bromo-9,9-dimethylfluorene). mp: 202 °C. 1H NMR (300 MHz,
CDCl3): δ 7.77-7.65 (m, 8H), 7.44 (m, 2H), 7.35 (m, 4H), 7.07-6.88
(m, 20H), 1.40 (s, 12H). 13C NMR (75 MHz, CDCl3): δ 156.08, 153.68,
152.86, 141.02, 139.95, 139.72, 138.81, 138.58, 134.70, 131.28, 130.50,
130.19, 129.93, 129.18, 127.65, 127.33, 126.89, 126.45, 126.27, 125.50,
122.58, 120.24, 119.82, 46.82, 26.97. HRMS (EI): m/z 770.3332 (calcd
for C58H46Si, 770.3369). Anal. Calcd for C58H46Si: C, 90.34; H, 6.01.
Found: C, 89.96; H, 5.95. UV (CHCl3), λmax (ꢀmax) 278 (5.38 × 104),
311 (5.49 × 104), 368 (0.76 × 104) nm (mol-1 L cm-1).
157.78, 151.00 (dm, JCF ) 246 Hz, C6F5), 144.89 (dm, JCF ) 259 Hz,
C6F5), 139.19 (dm, JCF ) 257 Hz, C6F5), 137.89, 136.68, 134.78, 129.57,
129.15, 128.01, 127.76, 127.11, 126.68, 103.76 (m, C6F5). HRMS
(EI): m/z 718.1191 (calcd for C40H20F10Si, 718.1175). Anal. Calcd for
C40H20F10Si: C, 66.85; H, 2.81; F, 26.44. Found: C, 66.77; H, 2.84;
F, 26.25. UV (CHCl3), λmax (ꢀmax) 239 (1.45 × 104), 376 (0.38 × 104)
nm (mol-1 L cm-1).
PES and IPES Data Collection. The PES and IPES measurements
were performed in a multichamber ultrahigh vacuum system comprising
a growth chamber with organic evaporation stations and quartz-crystal
flux monitor connected to a surface analysis chamber equipped with
PES and IPES. The four materials were loaded in separate PBN
crucibles in the growth chamber. The substrates were Si wafers covered
with a 50 Å Ti layer (for adhesion) and a 1200 Å Au layer. Prior to
insertion in a vacuum, the substrates were degreased by being boiled
in trichloroethylene and rinsed in acetone and methanol. Silole films
with thickness ranging from 20 to 50 Å were deposited on the Au
surface at a rate of 1 Å/s (assuming a density of 1.5 g/cm3) and
transferred to the analysis chamber. PES was performed using the
standard He I (21.22 eV) and He II (40.8 eV) photon lines from a He
discharge lamp and a cylindrical mirror analyzer, giving an experimental
resolution of 150 meV. IPES was carried out in the isochromat mode
using a fixed-photon-energy detector centered at 9.2 eV15 and a Kimbal
Physics electron gun, giving a combined resolution of 500 meV. The
IPES electron-beam current density was limited to 1 µA/cm2 to
minimize degradation of the organic material. In all cases, PES was
carried out first on the fresh film and repeated after the IPES
measurement to assess electron-beam-induced degradation. The PES
and IPES energy scales were aligned by measuring the position of the
Fermi level of a freshly evaporated Au film. The position of the vacuum
level at the surface of each film, Evac, was measured using the onset of
photoemission16 to determine IP and EA. In accord with the ad hoc
procedure widely accepted in the literature,17 the experimental values
of adiabatic IP and EA were taken as the energy difference between
Evac and the leading edge of the HOMO and LUMO feature,
respectively.
Computational Details. The geometries of I-V were optimized in
the neutral, radical-anion, and radical-cation states via density functional
theory (DFT). The DFT calculations were carried out using the B3LYP
functional, where Becke’s three-parameter hybrid exchange functional
is combined with the Lee-Yang-Parr correlation functional,18-20 with
a 6-31G* split valence plus polarization basis set. Excitation energies
for the low-lying excited states were calculated with time-dependent
density functional theory (TD-DFT). Simulation of the PES and IPES
spectra was accomplished through the density of states (DOS) given
by the DFT methodology. To account for polarization effects in the
solid state, the DOS was rigidly shifted with respect to the binding-
energy axis;21 the DOS was convoluted with Gaussian functions
characterized by a full width at half-maximum (fwhm) of ca. 0.5-0.7
eV to replicate the experimental line widths. Since the B3LYP
functional contains correlation effects, neither compression nor expan-
sion of the DOS was performed, in contrast to previous Hartree-Fock-
based simulations of PES and IPES data.22 All DFT calculations were
performed with Gaussian98 (revision A.11).23
1,1-Bis(thien-2-yl)-2,3,4,5-tetraphenylsilole (IV). To a 250-mL
three-necked round-bottom flask were added 2-bromothiophene (12.6
mmol, 2.06 g) and 20 mL of dry diethyl ether; the solution was
deoxygenated with nitrogen for 30 min and cooled to -78 °C. A
quantity of 7.9 mL of a 1.6 M hexane solution of n-butyllithium (12.6
mmol) was added dropwise. The mixture was stirred at the same
temperature for 1 h. The THF solution of VI was cooled to 0 °C and
transferred portionwise into 2-thienyllithium over a period of 30 min.
The brown mixture was stirred for 1 h at -78 °C, gradually warmed
to room temperature, and stirred overnight. The resulting dark yellow-
green solution was washed with water; the organic layer was extracted
with diethyl ether and dried over anhydrous magnesium sulfate. The
solvent was removed, and the residue was purified by flash column
chromatography over silica gel using hexane/dichloromethane ) 4/1
as eluent. Recrystallization from ethanol gave faintly greenish-yellow
and highly blue fluorescent crystals (1.4 g, 41%, based on silicon
1
tetrachloride). mp: 212 °C. H NMR (300 MHz, CDCl3): δ 7.71 (d,
J ) 4.20 Hz, 2H), 7.51 (d, J ) 3.30 Hz, 2H), 7.23 (dd, J ) 3.60, 3.30
Hz, 2H), 7.10-6.93 (m, 14H), 6.93-6.80 (m, 6H). 13C NMR (75 MHz,
CDCl3): δ 156.37, 138.74, 138.42, 133.10, 129.90, 129.61, 129.45,
128.54, 127.78, 127.47, 126.53, 125.88. HRMS (EI): m/z 550.1208
(calcd for C36H26S2Si, 550.1245). Anal. Calcd for C36H26S2Si: C, 78.50;
H, 4.76; S, 11.64. Found: C, 78.30; H, 4.83; S, 11.46. UV (CHCl3),
λmax (ꢀmax) 242 (4.03 × 104), 372 (0.80 × 104) nm (mol-1 L cm-1).
1,1-Bis(pentafluorophenyl)-2,3,4,5-tetraphenylsilole (V). To a 250-
mL three-necked round-bottom flask were added magnesium shavings
(25 mmol, 0.6 g), bromopentafluorobenzene (12.6 mmol, 3.11 g), and
a little iodine, and these were deoxygenated with nitrogen for 30 min.
A quantity of 30 mL of dry THF was added, and the mixture was stirred
at 40 °C for 2 h. The gray Grignard reagent suspension was cooled to
0 °C and transferred portionwise into the THF solution of VI over a
period of 30 min. The brown mixture was stirred for 1 h at 0 °C,
gradually warmed to room temperature, and stirred overnight. The
resulting dark yellow-green solution was washed with water, and the
organic layer was extracted with ether and dried over anhydrous MgSO4.
The solvent was removed, and the residue was purified by flash column
chromatography over silica gel using hexane/dichloromethane ) 2/1
as eluent and recrystallized from acetonitrile to afford yellow-green
and highly green fluorescent crystals (2.1 g, 46%, based on silicon
(15) Wu, C. I.; Hirose, Y.; Sirringhaus, H.; Kahn, A. Chem. Phys. Lett. 1997,
272, 43-47.
(16) Cahen, D.; Kahn, A. AdV. Mater. 2003, 15, 271-277.
(17) Shen, C.; Kahn, A.; Hill, I. G. In Conjugated polymer and molecular
interfaces: science and technology for photonic and optoelectronic
applications; Salaneck, W. R., Seki, K., Kahn, A., Pireaux, J. J., Eds.;
Marcel Dekker: New York, 2002.
(18) Becke, A. D. Phys. ReV. A 1988, 38, 3098-3100.
(19) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652.
(20) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785-789.
(21) The rigid shifts required for the fitting of the PES/IPES spectra using the
DFT DOS were as follows: II, 2.97 eV PES, 5.22 eV IPES; III, 2.04 eV
PES, 5.16 eV IPES; IV, 2.79 eV PES, 4.64 eV IPES; and V, 2.83 eV PES,
5.20 eV IPES.
1
tetrachloride). mp: 190 °C. H NMR (300 MHz, CDCl3): δ 7.12-
(22) Hill, I. G.; Kahn, A.; Cornil, J.; dos Santos, D. A.; Bre´das, J. L. Chem.
Phys. Lett. 2000, 317, 444-450.
6.96 (m, 12H), 6.94-6.80 (m, 8H). 13C NMR (75 MHz, CDCl3): δ
9
J. AM. CHEM. SOC. VOL. 127, NO. 25, 2005 9023

A R T I C L E S
Zhan et al.
Scheme 1. Synthetic Routes to Siloles III, IV, and V
tions neglect solid-state polarization effects; thus, the apparent
close similarity between calculated gas-phase and experimental
thin-film IP values is an adventitious artifact of the computa-
tional methodology used. The experimental IPs fall in the range
6.05-6.78 eV, considerably higher than the solid-state IP values
for typical hole-transport (HT) materials such as N,N′-diphenyl-
N,N′-di(3-methylphenyl)-(1,1′-biphenyl)-4,4′-diamine (TPD) (5.38
eV, adiabatic) and NPD (5.5 eV29,30), suggesting a high barrier
for hole injection into the siloles from these materials. The EAs
(-1.5 to -2.4 eV) may be compared to a value in the range
-2.0 to -2.5 eV measured for the widely used ET material,
Alq3.16 This suggests that electron-injection barriers similar to
those obtained with Alq3 could be realized with some of these
siloles. The experimental IP and EA values for II, III, and IV
are close, within experimental error, whereas V is harder to
oxidize and easier to reduce than II-IV. The DFT calculations
reproduce this pattern.
The calculated HOMOs and LUMOs (experimentally probed
by the PES and IPES measurements, respectively) for II-V
do not vary drastically in appearance among the different
compounds; moreover, they are similar to those reported in other
computational studies of siloles.10,11,31 The HOMO is similar
to the Hu¨ckel HOMO for cisoid-butadiene, with some additional
antibonding contributions from the local HOMOs of 2,5- and,
to a lesser extent, 3,4-phenyl groups; see Figure 3. The LUMO
resembles the LUMO of butadiene with additional in-phase
contributions from a Si-Ar σ* orbital and from the local
LUMOs of the 2,3,4,5-phenyl groups. Significantly, neither
HOMO nor LUMO shows any obvious contributions from the
π-orbitals of the 1,1-diaryl groups; thus, the variation of the
substituents on silicon appears to affect the electronic structure
through a principally inductive mechanism. While this has been
proposed in previous studies,11 it has also been suggested that
when the 1,1-substituents are phenyl, π-effects are important.11
However, support for a mainly inductive role of the aryl groups
comes from the similar effects of a given substituent on both
HOMO and LUMO energies (see Table S2 in Supporting
Information). Moreover, the computed IP and EA data suggest
slightly lower-lying HOMO and LUMO for the thienyl species,
IV; although thiophene is an electron-rich π-system, it has
previously been shown that the 2-thienyl group is inductively
somewhat electron-withdrawing.32
III. Results and Discussion
Synthesis. 1,1-Dimethyl-2,3,4,5-tetraphenylsilole (I) and 1,1-
diphenyl-2,3,4,5-tetraphenylsilole (II) were prepared according
to published procedures8 from the ring-closing reaction of 1,4-
dilithio-1,2,3,4-tetraphenyl-1,3-butadiene with dichlorodimeth-
ylsilane or dichlorodiphenylsilane, respectively. The new com-
pounds III-V were made by quenching the same organolithium
reagent with silicon tetrachloride to afford the known 1,1-
dichlorosilole intermediate, VI. VI was then transformed to
siloles III, IV, and V by the reactions with Grignard or lithium
reagents, as shown in Scheme 1. All siloles were prepared in
40-50% yields. All silole products were fully characterized by
spectroscopic methods and elemental analysis.
IP and EA Measurements and Simulations. The most direct
probes of the energies of filled and empty orbitals of molecules
are PES and IPES, respectively. Data were acquired on vapor-
deposited thin films of the 1,1-diaryl-substituted siloles, II-
V,24 since these data are more relevant to device applications.
It should be noted that experimental thin-film IPs for small
molecules have previously been shown to be as much as 1-1.5
eV lower than the experimental gas-phase values due to extra
stabilization of the cation in the solid state through polarization
of the surrounding medium;25-27 accordingly, EA values are
expected to be more exothermic in the solid state than the gas
phase due to polarization stabilization of the anion.28 The spectra
are shown in Figure 2, along with simulations based upon DFT
calculations that facilitate interpretation and assignment of the
spectra; it should be noted that few such comparisons between
theoretical and experimental densities of states of both filled
and empty orbital structures have been published.22 The
experimentally determined IPs and EAs are given in Table 1.
These estimates of the adiabatic values are obtained from the
onsets of the spectra, the relatively large error bars at least
partially stemming from difficulties associated with accurate
determinations of these onsets. The adiabatic and vertical gas-
phase IPs/EAs from DFT self-consistent-field calculations are
included for comparison. The absolute experimental values are
not expected to agree with those calculated, since the calcula-
Electrochemistry. Since cyclic voltammetry is a much more
routine experiment than either PES or IPES, we were interested
in establishing the extent to which the trends seen in PES and
IPES, and in the calculations, could be discerned from electro-
chemical data. Moreover, it is of interest to compare how IP
and EA energies estimated from the electrochemical data
compare with those determined more directly by PES and IPES.
An example of a cyclic voltammogram is illustrated in Figure
4. The siloles I-V show one (II-V) or two (I) oxidation peaks
and two reduction peaks, none of which appears fully reversible,
indicating the charged species to be unstable in solution; the
anodic peak potentials, Epa, and cathodic peak potentials, Epc,
are summarized in Table 2. Due to the incomplete reversibility
of these processes, the peaks do not necessarily represent the
(23) Frisch, M. J. et al. Gaussian98, revision A.11; Gaussian, Inc.: Pittsburgh,
PA, 1998.
(24) We were unable to acquire spectra for I for comparison since I was too
volatile under the high-vacuum conditions required for the spectroscopy.
(25) Anderson, J. D. et al. J. Am. Chem. Soc. 1998, 120, 9646-9655.
(26) Sato, N.; Inokuchi, H.; Silinsh, E. A. Chem. Phys. 1987, 115, 269-277.
(27) Silinsh, E. A.; Capek, V. Organic Molecular Crystals: Interaction,
Localization, and Transport Phenomena; AIP Press: New York, 1994.
(28) Interestingly, computationally estimated polarization energies for anthracene
are similar for both holes and electrons, whereas for perylene tetracarboxylic
acid dianhydride they differ by more than 2 eV (Tsiper, E. V.; Soos, Z. G.
Phys. ReV. B 2001, 64, 195124-1-195124-12).
(29) Kahn, A.; Koch, N.; Gao, W. Y. J. Polym. Sci., Part B: Polym. Phys.
2003, 41, 2529-2548.
(30) Gao, W. Y.; Kahn, A. J. Appl. Phys. 2003, 94, 359-366.
(31) Risko, C.; Kushto, G. P.; Kafafi, Z. H.; Bre´das, J. L. J. Chem. Phys. 2004,
121, 9031-9038.
(32) Fringuel, F.; Marino, G.; Taticchi, A. J. Chem. Soc. B 1971, 2302-2303.
9
9024 J. AM. CHEM. SOC. VOL. 127, NO. 25, 2005

EAs of 1,1-Diaryl-2,3,4,5-tetraphenylsiloles
A R T I C L E S
Figure 2. Comparison between PES and IPES spectra measured for II-V. For each compound, the thin-film (top) and DFT-simulated (bottom) spectra are
given; the vertical bars refer to the shifted energies of the molecular orbitals.
Table 1. Experimental PES and IPES Estimates of Thin-Film
Adiabatic IP and EA (eV, from Onset) and DFT Calculations of
Gas-Phase Adiabatic and Vertical IP and EA
IP
EA
PES
DFT
adiabatic
IPES
DFT
adiabatic
vertical
vertical
I
II
-
6.30
6.19
6.10
6.23
6.49
6.54
-
-0.62 -0.34
6.19 ( 0.10
6.42 -1.85 ( 0.40 -0.78 -0.52
6.29 -1.49 ( 0.40 -0.83 -0.58
6.46 -1.93 ( 0.40 -0.85 -0.58
6.85 -2.41 ( 0.40 -1.16 -0.86
III 6.40 ( 0.10
IV
V
6.05 ( 0.10
6.78 ( 0.10
thermodynamic oxidation and reduction potentials and should
be interpreted with caution. However, the electrochemical data
do reveal a picture similar to that of the PES and IPES data:
II-IV are similar to one another in IP and EA, with V both
significantly easier to reduce and harder to oxidize than I-IV.
The differences in the electrochemical data are less pronounced
than those in the experimental and calculated IP and EA data,
presumably due to differences in solvation effects. These results
are consistent with previous cyclic voltammetry studies upon
the influence of 1,1-substituents in which Dhiman and co-
workers reported that siloles with electron-withdrawing 1,1-
substituents (e.g., chlorine) show higher first oxidation poten-
tials.9
Figure 3. B3LYP/6-31G*-calculated highest occupied (HOMO) (left) and
lowest unoccupied (LUMO) (right) one-electron molecular orbitals for II.
determined by PES (e.g., the commonly used hole-transport
molecule TPD). From the electrochemical data, one would then
estimate the IP of I-V to be in the range ca. 6.0-6.2 eV (ca.
6.1 eV for II) using the relation
IP(M) - IP(TPD) ) E1/2(M+/0) - E1/2(TPD+/0
)
with values of 5.38 eV for the IP of TPD (thin film) and +0.35
V for E1/2(TPD+/0) (in MeCN/benzene),25 and assuming E1/2 is
similar to Epa1. These values are of similar magnitude to those
from PES (6.05-6.78 eV), although the range of values seen
in PES is considerably reduced in the electrochemical data and
the value for V is significantly underestimated.
The solid-state IP of a molecular species is often estimated
from comparison of the electrochemical potential for the M+/0
couple with that for a compound whose solid-state IP has been
9
J. AM. CHEM. SOC. VOL. 127, NO. 25, 2005 9025

A R T I C L E S
Zhan et al.
Table 3. B3LYP/6-31G* Intramolecular Reorganization Energies
as Determined by ∆SCFa
neutral/cation
2 (eV)
neutral/anion
2 (eV)
λ
1 (eV)
λ
λtotal (eV)
λ1 (eV)
λ
λtotal (eV)
Ib
II
III
IV
V
0.21
0.20
0.19
0.21
0.25
0.24
0.22
0.19
0.23
0.36
0.45
0.42
0.38
0.44
0.61
0.24 0.22
0.23
0.24
0.24
0.26
0.27 0.28
0.26
0.26
0.27
0.30
0.51 0.50
0.49
0.50
0.51
0.56
a The total intramolecular reorganization energy (λtotal) combines the
relaxation energies of the initially neutral molecule (λ1) and the initially
ionized molecule (λ2) upon electron transfer (polaron hopping). b For I,
values in italics were determined by vibrational mode analysis.
and in that for Alq3 preclude detailed comparison of the
electrochemically estimated and IPES values of EA, it does seem
that EA energies are best estimated from comparison of
Figure 4. Cyclic voltammogram of silole III (with FeCp2 as an internal
reference) in CH3CN/toluene (1:1 volume)/0.1 M [nBu4N]+[PF6]- at 50
mVs-1. The horizontal scale refers to an anodized Ag wire pseudo-reference
electrode.
E
1/2(M0/-) values with that of a compound of known EA,
Table 2. Electrochemical Data (V) for I-Va
whereas IP energies are best estimated from comparison of
E
1/2(M+/0) values with that of a compound of known IP.
b
b
b
b
Epa1
Epa2
Epc1
Epc2
Reorganization Energies. Since siloles can show moderate
I
+0.95
+1.02
+1.07
+1.05
+1.17
+1.10
-
-2.43
-2.35
-2.41
-2.25
-2.08
-2.84
-2.57
-2.60
-2.48
-2.40
electron mobilities in excess of that seen for Alq3, we were
interested in comparing the barriers to electron transfer between
anionic and neutral species in the two systems; according to
Marcus theory,33-35 the barrier height, ∆Gq, for a self-exchange
reaction is given by λ/4, where λ is the reorganization energy
associated with the reaction:
II
III
IV
V
-
-
-
a In CH3CN/toluene (1:1 volume)/0.1 M [nBu4N]+[PF6]-, versus ferro-
cenium/ferrocene at 50 mVs-1 b Epa1 and Epa2 are the peak potentials
corresponding to successive molecular oxidations, while Epc1 and Epc2 are
.
the peak potentials corresponding to successive molecular reductions.
MA + MB- ) MA- + MB
EAs can, in principle, be estimated according to:
The reorganization energy (λ) can be separated into the sum
of two primary components: (i) the medium reorganization
energy that arises from modifications to the polarization of the
surrounding medium due to the presence of excess charge (λs)
and (ii) the intramolecular reorganization energy (λi), which
combines the relaxation energies of the electron-donor molecule
(λ1) and electron-acceptor molecule (λ2) associated with the
electron-transfer reaction. We calculated the intramolecular
contribution to the total reorganization energy for electron
transfer and the corresponding value for hole transfer from the
adiabatic potential surfaces for siloles I-V; the results are
compared in Table 3. The total reorganization energies for
electron transfer are ca. 0.5 eV, values comparable with those
previously obtained for 1 and 2 at the same level of theory,31
while those for hole transfer are in the range 0.4-0.6 eV. The
EA(M) + IP(TPD) ) E1/2(TPD+/0) - E1/2(M0/-
)
For II, a value of ca. -2.7 eV is estimated, suggesting injection
of an electron into the solid to be almost 1 eV easier than is
indicated by the IPES data. This poor agreement with the IPES
value can be attributed to different magnitudes of solvation
effects and polarization effects on E1/2 and IP/EA, respectively;
solvation appears to stabilize TPD+ and/or M- more than solid-
state effects. An alternative estimate of the EA that is often
used is to combine the IP and the optical gap, Eop (estimated
0-0 transition, vide infra), according to:
EA ) Eop - IP
For II, a value of ca. -3.2 to -3.1 eV is obtained (using
either the PES or the electrochemical estimate of IP), also more
than 1 eV different from the IPES value. This method has also
been previously used to estimate similar values for the EA of
siloles (1, 2, 4) from PES-determined IPs.13 The poor agreement
with the IPES value of EA using this method is largely
attributable to the neglect of the exciton-binding energy (vide
infra).
reorganization energy values for electron transfer are ap-
0/- 36
proximately twice that calculated at the same level for Alq3
,
while those for hole transfer are approximately four to five times
those calculated for [pentacene]+/0 37 and nearly twice those
calculated for TPD+/0 38,39
Thus, the differences in electron
.
mobilities that have been seen between siloles and Alq3 are
likely to be due to differences in intermolecular orbital overlap
or to morphological effects (such as trapping at grain bound-
An alternative electrochemical estimate of EA can be obtained
from comparison to a material of known EA and reduction
potential, such as Alq3, via:
(33) Marcus, R. A. J. Chem. Phys. 1956, 24, 966.
(34) Barbara, P. F.; Meyer, T. J.; Ratner, M. A. J. Phys. Chem. 1996, 100,
13148-13168.
EA(M) - EA(Alq3) ) E1/2(Alq30/-) - E1/2(M0/-
)
(35) Bre´das, J. L.; Beljonne, D.; Coropceanu, V.; Cornil, J. Chem. ReV. 2004,
104, 4971-5003.
(36) Lin, B. C.; Cheng, C. P.; You, Z.-Q.; Hsu, C.-P. J. Am. Chem. Soc. 2005,
127, 66-67.
Taking E1/2(Alq30/-) ) -2.30 V,25 this method suggests that
the EA for II is close to that for Alq3 (-2.0 to -2.5 eV16),
closer to the values from IPES data than those estimated above.
While the large uncertainties in the present IPES EA values
(37) Gruhn, N. E.; da Silva, D. A.; Bill, T. G.; Malagoli, M.; Coropceanu, V.;
Kahn, A.; Bre´das, J. L. J. Am. Chem. Soc. 2002, 124, 7918-7919.
(38) Malagoli, M.; Bre´das, J. L. Chem. Phys. Lett. 2000, 327, 13-17.
(39) Lin, B. C.; Cheng, C. P.; Lao, Z. P. M. J. Phys. Chem. A 2003, 107, 5241-
5251.
9
9026 J. AM. CHEM. SOC. VOL. 127, NO. 25, 2005

EAs of 1,1-Diaryl-2,3,4,5-tetraphenylsiloles
A R T I C L E S
Table 4. Representative Vibrational Modes Corresponding to the
Relaxation in the Neutral and Radical-Anion Potential Surfaces of
I′ (the Energy Value Represents the Contribution of Each Mode to
the Total Relaxation Energy)
and it is thought that restriction of such rotation is responsible
for the aggregation-induced emission effects often observed for
substituted silole molecules.42,43 In the radical-anion potential
well, 65% of the total relaxation energy is composed of six
vibrational modes. The majority of the relaxation energy
contribution (56%) comes from various ring-breathing mode
combinations of the silole and phenyl rings, with significant
contribution (9%) coming from phenyl-ring rotation with respect
to the central silole ring. Thus, for this model system, the
majority of the total intramolecular reorganization energy arises
from combinations of ring-breathing modes of the central silole
ring and the 2,5-susbtituted phenyl rings. Normal-mode analysis
of I0/- reveals a picture similar to that provided by I′0/-, though
the complexity of the modes involved is much more significant
due to the number of intrinsic degrees of freedom in the
molecule. Again, λ obtained by the normal-mode analysis (0.50
eV) is close to that obtained via the adiabatic potential surface
analysis (0.51 eV). Here, upward of 15 vibrational modes play
significant roles in the relaxation processes on the neutral and
radical-anion surfaces (see Supporting Information Table S5).
As with I′0/-, the predominant vibrations consist of various ring-
breathing modes and phenyl ring torsions, as well as Me-Si
stretches/bends.
UV-Vis Absorption and Fluorescence Spectroscopy.
Solution (chloroform) and thin-film UV-vis absorption spectra
for the silole systems are shown in Figure 5, with the spectral
characteristics summarized in Table 5. The lowest-energy
maxima are essentially independent of the identity of the aryl
group and little shifted between the solution and solid states,
consistent with previous observations for other siloles.44,45 TD-
DFT calculations indicate that these transitions can be described
as predominantly HOMOfLUMO in character (i.e., π-π*
silacyclopentadiene transitions), but they somewhat underesti-
mate the transition energies.46
Solution (chloroform) and thin-film PL spectra of the 1,1-
substituted siloles I-V are shown in Figure 5, with emission
data provided in Table 5. All five siloles show weak blue-green
luminescence in solution, with very similar emission spectra
peaked around 2.51-2.61 eV. Consistent with other siloles,42,43
the fluorescence appears to be considerably brighter in the solid
state relative to solution; however, the maxima are seen at
similar energies (the features seen in the solid-state bands are
slightly narrower, but this is presumably attributable to self-
absorption).47
From the magnitude of the bulk transport gap Et, also known
as the single-particle gap, and the optical properties, one can
estimate the solid-state exciton-binding energy (EB). This is the
energy required for photocharge generation of a silole anion
and silole cation from an excited silole molecule (Frenkel
neutral
ω
(cm-
energy
(eV)
1
)
S
mode description
neutral
anion
29 12.2
0.044 rotation of phenyl rings
182
947
0.58 0.013 methyl-silicon-methyl bend
0.12 0.014 ring breathing of silole & phenyl rings
0.20 0.038 ring breathing of silole & phenyl rings
0.40 0.077 ring breathing of silole & phenyl rings
8.83 0.023 rotation of phenyl rings
1536
1550
21
934
0.12 0.014 ring breathing of silole & phenyl rings
0.19 0.030 ring breathing of silole & phenyl rings
0.22 0.040 ring breathing of silole & phenyl rings
0.15 0.028 ring breathing of silole & phenyl rings
0.07 0.014 ring breathing of silole & phenyl rings
0.07 0.014 ring breathing of silole & phenyl rings
1250
1485
1494
1537
1646
aries), rather than to differences in λ. The geometric modifica-
tions that occur on both reduction and oxidation of the silole
systems are primarily localized toward the central portion of
the molecular structures, particularly in the silole ring and in
the exocyclic Si-Ar bonds, but also to some extent in some of
the carbon-carbon bonds of the 2,5-phenyl rings in close
proximity to the silole ring and in the torsional angle of the
2,5-phenyl rings with respect to the silole ring (see Tables S3
and S4 and Figure S1 in Supporting Information).40 This
localization is mainly responsible for the relatively large
reorganization energy.
To gain further insight into the reorganization energy associ-
ated with the electron-exchange reaction between neutral and
anionic siloles, we explored the contribution of different
vibrational modes to the reorganization energy. Since the
reorganization energies do not depend strongly upon the
substituents on the silicon atom and since the geometric changes
do not extend into this substituent, we investigated the simplest
system, I; in addition, we also examined the model compound
I′, in which the 3,4-phenyl groups are replaced by hydrogens,
to reduce the number of degrees of freedom and simplify
interpretation. The contribution of normal modes to the reor-
ganization energy can be obtained through the relation:
λtotal
)
λ )
S pω
i i
∑ ∑
i
where S is the Huang-Rhys factor.34,41 Normal-mode analysis
for I′0/- gives a value of 0.51 eV for λ, in good agreement with
the value (0.47 eV) obtained from the adiabatic potential
surfaces, and of similar magnitude to the values in Table 3 for
I-V.51 Five vibrational modes in I′0 within the neutral potential
well comprise 73% of the total relaxation energy, as shown in
Table 4. Of these five modes, three ring-breathing modes
provide 51% of the total relaxation energy, while rotation of
the phenyl rings with respect to the central silole ring makes
up 17% of the total relaxation energy. Previous experimental
evidence from a variety of techniques (solution fluorescence
spectroscopy with variable temperatures and solvent viscosities,
along with variable-temperature NMR experiments) suggests
that such a rotational mode plays an important role in the
intensity dampening of singlet emission for siloles in solution,
(42) Luo, J. D. et al. Chem. Commun. 2001, 1740-1741.
(43) Chen, J. W.; Law, C. C. W.; Lam, J. W. Y.; Dong, Y. P.; Lo, S. M. F.;
Williams, I. D.; Zhu, D. B.; Tang, B. Z. Chem. Mater. 2003, 15, 1535-
1546.
(44) It should also be noted that spectra in nonpolar solvents (hexane or toluene)
showed almost identical absorption maxima, as has been previously reported
for other silole systems; see ref 10.
(45) We attribute the sharp absorption peaks in III to π-π* transitions within
the fluorenyl groups.
(46) The calculations do not successfully reproduce the slight variations in
absorption maxima between the compounds; in particular, the experimental
maximum for the pentafluorophenyl derivative, V, is slightly red-shifted
relative to that of the phenyl compound, II, but is calculated to be slightly
blue-shifted. These differences may be due to experimental maxima being
somewhat shifted from the vertical transition energies due to overlap with
the tails of the higher energy (Ph and other aryl π-π*) transitions.
(40) These torsions are likely suppressed in the solid state.
(41) Pope, M.; Swenberg, C. E. Electronic Processes in Organic Crystals and
Polymers, 2nd ed.; Oxford University Press: New York, 1999.
9
J. AM. CHEM. SOC. VOL. 127, NO. 25, 2005 9027

A R T I C L E S
Zhan et al.
Figure 5. Solution (chloroform) (top left) and thin-film (top right) UV-vis absorption spectra and solution (chloroform) (bottom left) and thin-film (bottom
right) fluorescence spectra of I-V.
Table 5. Experimental and TD-DFT Absorption and Experimental
Emission Characteristics for the Lowest-Lying Excited States of
I-V
Table 6. Exciton-Binding Energies as Determined by PES/IPES
and DFT Calculated Ionization Potentials and Electron Affinities,
and Solid-State UV/Vis and TD-DFT Absorption Data
absorption (eV)a
emission (eV)a
method
IP (eV)
EA (eV)a
E
op (eV)b
Et (eV)c
EB (eV)
experiment
TD-DFT
PES/IPES
I
-
-
-
-
II
III
IV
V
6.19 -1.85 2.97 3.36
6.40 -1.49 2.94 3.38
6.05 -1.93 2.94 3.31
6.78 -2.41 2.91 3.28
6.30 -0.62 3.23
4.54
5.11
4.32
4.57
5.88
5.0
1.57 1.18
2.17 1.73
1.38 1.01
1.66 1.29
2.65
I
3.43 (3.41)
3.38 (3.36)
3.37 (3.38)
3.33 (3.31)
3.30 (3.28)
3.23
3.10
3.08
3.05
3.14
2.61 (2.54)
2.54 (2.53)
2.53 (2.51)
2.51 (2.52)
2.53 (2.46)
II
III
IV
V
DFT/TD-DFT
I
II
III
IV
V
6.19 -0.78 3.10
1.90
6.10 -0.83 3.09
5.9
2.81
a Measured in chloroform; data in thin films are given in the parentheses.
exciton), that is, the energy required for the reaction:
6.23 -0.85 3.04
4.6
5.25
1.56
6.49 -1.16 3.14
2.11
a Note that the electron affinities are calculated by subtracting the total
energy on the neutral electronic configuration from the energy of the radical-
anion electronic configuration. b Experimental values for Eop estimated from
the midpoint of absorption and fluorescence and from the absorption maxima
(italics), for direct comparison with the DFT/TD-DFT results. c Experimental
values of the transport gap are obtained as explained in the text.
M* + M ) M+ + M-
which is given by:
EB ) Et - Eop
increase in polarization from surface to bulk (0.3 eV for the
hole states and 0.3 eV for the electron states) and (ii) the loss
of energy of the photoemitted electron (PES) and injected
electron (IPES) to molecular vibration (2 × 0.1 eV; i.e.,
providing an approximate correction from vertical to adiabatic
values). The experimental values shown in Table 6 indicate
exciton-binding energies in the range 1.0-1.7 eV when Eop is
estimated from the absorption maximum, similar to those
estimated by the same method for other organic systems.48,49
Values obtained from the DFT adiabatic surface energies and
from TD-DFT absorption energies are much larger (ca. 2 eV),
where Eop is the 0-0 maximum. We evaluated Et using a
procedure previously established by Hill et al.48 Et is taken as
the experimental peak-to-peak HOMO-to-LUMO gap measured
by PES and IPES, reduced by 0.8 eV to account for: (i) the
(47) The fact that strong luminescence is observed in the solid state, along with
minimal shifts in the luminescence energy, is consistent with previous
reports for siloles. These results are in contrast to “typical” organic emitters
that tend to undergo various degrees of aggregation and excimer formation
that leads to concentration-dependent electroluminescence quenching,
emission band broadening, and bathochromic shifts. The fact that siloles
do not undergo these processes has been explained by the quenching of
nonradiative decay routes in the solid state; see ref 14.
(48) Hill, I.; Kahn, A.; Soos, Z.; Pascal, R. Chem. Phys. Lett. 2000, 327, 181-
188.
(49) Knupfer, M. Appl. Phys. A 2003, 77, 623-626.
9
9028 J. AM. CHEM. SOC. VOL. 127, NO. 25, 2005

EAs of 1,1-Diaryl-2,3,4,5-tetraphenylsiloles
A R T I C L E S
though with a similar variation in energy across the series (0.25
eV); the discrepancy between experimental and DFT estimates
of EB can be attributed at least partly to the neglect in the DFT
approach of the solid-state polarization effects that stabilize both
cations and anions relative to the gas phase.
We have also used DFT to calculate the reorganization energy
associated with the electron-transfer reactions of siloles and their
radical cations or anions. These reorganization energies are
somewhat larger than those for Alq3, suggesting that this is not
the factor determining the higher electron mobilities reported
in siloles. From optical data and the transport gap deduced from
PES and IPES data, we have been able to estimate exciton-
binding energies for siloles. Finally, the similarity of the EAs
we have measured to that of the widely used Alq3, combined
with the high electron mobilities previously reported for other
siloles, suggests that these materials could potentially replace
Alq3 in many ET applications; the C6F5 derivative, which has
the lowest-lying LUMO and, therefore, the lowest barrier to
electron injection, is particularly promising.
IV. Conclusion
We have measured the solid-state EAs of siloles using IPES;
these are the most direct measurements to date of these important
parameters for gauging their compatibility with other organic
electronic materials and with inorganic electrode materials. Of
the functional groups we have examined, only the very strongly
electron-withdrawing pentafluorophenyl groups have a signifi-
cant effect on the IP and EA. The variations in the EA and IP
of 1,1-diaryl-2,3,4,5-tetraphenylsiloles with the identity of the
aryl groups, Ar, appear to be largely dictated by inductive
effects. This conclusion is also supported by DFT calculations.
Thus, thiophene, despite having an electron-rich π-system, acts
as a weakly electron-withdrawing group on the silole skeleton
when attached to the silicon atom. Another result of the inductive
role played by the Ar group is that the HOMO-LUMO gap
and, hence, the optical absorption or fluorescence properties are
rather insensitive to the identity of Ar.
Since EA measurements of ET materials are scarce, many
workers have estimated EA from electrochemical data, often
using the known IP or EA of a reference material.50 We have
shown here that estimates of the EAs of siloles from electro-
chemical data are only in good agreement with the IPES values
if one uses another ET material of known EA, such as Alq3, as
a reference material; estimates based on the known IP of TPD
lead to overestimation of the magnitude of the EAs by more
than 1 eV. Estimates of EA based on IP and optical data are
also poor since they neglect the exciton-binding energy.
Acknowledgment. This material is based upon work sup-
ported in part by the National Science Foundation STC Program
under Agreement No. DMR-0120967. We also thank the NSF
for support through Grants DMR-0408589, CHE-0342321, and
CHE-0211419 and the Office of Naval Research for support
through a MURI (Award No. N00014-03-1-0793, administered
through the California Institute of Technology). Support from
Georgia Institute of Technology, through the Center for Organic
Photonics and Electronics, from the New Jersey Center for
Organic Optoelectronics, and from the Princeton Institute for
the Science and Technology of Materials is also gratefully
acknowledged.
Supporting Information Available: Tables of thermal data
for I-V; calculated frontier orbital energies for I-V; calculated
geometric parameters for I-V in neutral, anionic, and cationic
states; calculated vibrational modes associated with the relax-
ation in neutral and anionic surfaces of I. Complete citations
for refs 23, 25, and 42. This material is available free of charge
via the Internet at http://pubs.acs.org.
(50) Domercq, B.; Grasso, C.; Maldonado, J. L.; Halik, M.; Barlow, S.; Marder,
S. R.; Kippelen, B. J. Phys. Chem. B 2004, 108, 8647-8651.
(51) Analysis of the vibrational mode contribution to the intramolecular
reorganization energy was carried out with the DUSHIN program of J. R.
Reimers (Reimers, J. R. J. Chem. Phys. 2001, 115, 9103).
JA051139I
9
J. AM. CHEM. SOC. VOL. 127, NO. 25, 2005 9029