Elastic and inelastic electron tunneling spectroscopy of a new rectifying monolayer

Article abstract of DOI:10.1021/ja068729g

Rectification of electrical current was observed in a Langmuir-Schaefer monolayer of fullerene-bis[ethylthio-tetrakis(3,4-dibutyl-2-thiophene-5-ethenyl) -5-bromo-3,4-dibutyl-2-thiophene] malonate, Au electrodes at room temperature (there are two regimes of asymmetry, at lower bias, i.e., between 0 and ±2 V, and at higher bias), and also between Pb and Al electrodes at 4.2 K. The latter experiment was coupled with second harmonic detection of the second derivative of the current with respect to voltage (d²I/dV ²). The d²I/dV² spectrum shows intramolecular vibrations, and also two antisymmetric broad bands, centered at ±0.65 V, due to resonant electron tunneling between the Fermi level(s) of the electrodes and the lowest unoccupied molecular orbital of the molecule.

Full text of DOI:10.1021/ja068729g

Published on Web 06/07/2007
Elastic and Inelastic Electron Tunneling Spectroscopy of a
New Rectifying Monolayer
Andrei Honciuc,†,‡ Robert M. Metzger,*,† Aijun Gong,§, and Charles W. Spangler§,|
Contribution from the Laboratory for Molecular Electronics, Chemistry Department, UniVersity
of Alabama, Tuscaloosa, Alabama 35487-0336, and Department of Chemistry, Montana State
UniVersity, Bozeman, Montana 59717
Received December 5, 2006; E-mail: rmetzger@ua.edu
Abstract: Rectification of electrical current was observed in a Langmuir-Schaefer monolayer of fullerene-
bis[ethylthio-tetrakis(3,4-dibutyl-2-thiophene-5-ethenyl)-5-bromo-3,4-dibutyl-2-thiophene] malonate, Au elec-
trodes at room temperature (there are two regimes of asymmetry, at lower bias, i.e., between 0 and (2 V,
and at higher bias), and also between Pb and Al electrodes at 4.2 K. The latter experiment was coupled
with second harmonic detection of the second derivative of the current with respect to voltage (d2I/dV2).
The d2I/dV2 spectrum shows intramolecular vibrations, and also two antisymmetric broad bands, centered
at (0.65 V, due to resonant electron tunneling between the Fermi level(s) of the electrodes and the lowest
unoccupied molecular orbital of the molecule.
Introduction
by scanning tunneling microscopy have yielded several other
rectifiers.12-16 Most of these studies coupled the measured I-V
Electrical rectification in “metal|organic monolayer|metal”
(MOM) junctions is a fundamental problem of asymmetric
electron transfer across one “electroactive” molecule (or many
molecules in parallel), but may also lead to single organic
molecules as components for future nanoscopic organic elec-
tronic devices.
The first proposed mechanism for electrical rectification by
a single molecule was that of Aviram and Ratner:1 At different
values of forward and reverse bias, the molecular energy levels
of the donor and acceptor are brought differently into resonance
with the Fermi level of the (macroscopic) metal electrodes, and
the current-voltage plot (I-V curve) should be asymmetric with
respect to positive and negative bias.2 Many unimolecular
rectifiers have been studied;3-11 single-molecule measurements
curves with a theoretical and ancillary spectroscopic character-
ization of these electroactive molecules, in solution, in bulk,
and in monolayers. The purpose was to associate the highest
occupied molecular orbital (HOMO), the lowest unoccupied
molecular orbital (LUMO), and the electrode Fermi energies
with the onset of enhanced electron current beyond a critical
forward bias (“turn-on voltage”). The mechanism of electron
transport in molecules and molecular wires is not completely
elucidated. Spectroscopic studies of the monolayer under
electrical bias are severely limited by the difficulty of probing
the molecules within the MOM junction by sensitive spectro-
scopic methods.17,18
Theoretical calculations provide excellent insights19-22 into
the electrical conduction through a molecule by Landauer’s
† University of Alabama.
(9) Honciuc, A.; Jaiswal, A.; Gong, A.; Ashworth, K.; Spangler, C. W.;
Peterson, I. R.; Dalton, L. R.; Metzger, R. M. J. Phys. Chem. B 2005, 109,
857-871.
‡ Present address: Department of Chemical and Biological Engineering,
University of Colorado, Boulder, CO 80309-0424.
§ Montana State University.
(10) Shumate, W. J.; Mattern, D. L.; Jaiswal, A.; Dixon, D. A.; White, T. R.;
Burgess, J.; Honciuc, A.; Metzger, R. M. J. Phys. Chem. B 2006, 110,
11146-11159.
Present address: Department of Chemistry, University of Rochester,
Rochester, NY 14627.
(11) Oleynik, I. I.; Kozhushner, M. A.; Posvyanskii, V. S.; Yu, L. Phys. ReV.
Lett. 2006, 96, 09803-1-09803-4.
| Present address: Rastris, 2100 Fairway Dr., Suite 104, Bozeman, MO
58715.
(12) Ashwell, G. J.; Tyrrell, W. D.; Whittam, A. J. J. Mater. Chem. 2003, 13,
2855-2857.
(1) Aviram, A.; Ratner, M. Chem. Phys. Lett. 1974, 29, 277-283.
(2) Stokbro, K.; Taylor, J.; Brandbyge, M. J. Am. Chem. Soc. 2003, 125, 3674-
3675.
(13) Ashwell, G. J.; Chwialkowska, A.; High, L. R. H. J. Mater. Chem. 2004,
14, 2848-2851.
(3) Metzger, R. M. Chem. Phys. 2006, 326, 176-187.
(4) Metzger, R. M.; Chen, B.; Ho¨pfner, U.; Lakshmikantham, M. V.; Vuillaume,
D.; Kawai, T.; Wu, X.; Tachibana, H.; Hughes, T. V.; Sakurai, H.; Baldwin,
J. W.; Hosch, C.; Cava, M. P.; Brehmer, L.; Ashwell, G. J. J. Am. Chem.
Soc. 1997, 119, 10455-10466.
(14) Ashwell, G. J.; Berry, M. J. Mater. Chem. 2005, 15, 108-110.
(15) Ashwell, G. J.; Tyrrell, W. D.; Whittam, A. J. J. Am. Chem. Soc. 2005,
126, 7102-7110.
(16) Morales, G. M.; Jiang, P.; Yuan, S.; Lee, Y.; Sanchez, A.; You, W.; Yu,
L. J. Am. Chem. Soc. 2005, 127, 10456-10457.
(5) Metzger, R. M.; Xu, T.; Peterson, I. R. J. Phys. Chem. B 2001, 105, 7280-
(17) Nowak, A. M.; McCreery, R. L. J. Am. Chem. Soc. 2004, 126, 16621-
7290.
16631.
(6) Honciuc, A.; Otsuka, A.; Wang, Y.-H.; McElwee, S. K.; Woski, S. A.;
Saito, G.; Metzger, R. M. J. Phys. Chem. B 2006, 110, 15085-15093.
(7) Baldwin, J. W.; Amaresh, R. R.; Peterson, I. R.; Shumate, W. J.; Cava, M.
P.; Amiri, M. A.; Hamilton, R.; Ashwell, G. J.; Metzger, R. M. J. Phys.
Chem. B 2002, 106, 12158-12164.
(8) Metzger, R. M.; Baldwin, J. W.; Shumate, W. J.; Peterson, I. R.; Mani, P.;
Mankey, G. J.; Morris, T.; Szulczewski, G.; Bosi, S.; Prato, M.; Comito,
A.; Rubin, Y. J. Phys. Chem. B 2003, 107, 1021-1027.
(18) Jaiswal, A.; Tavakoli, C. G.; Zou, S. Anal. Chem. 2006, 78, 120-124.
(19) Krzeminski, C.; Delerue, C.; Allan, G.; Vuillaume, D.; Metzger, R. M.
Phys. ReV. 2001, B64, #085405.
(20) Gonzalez, C.; Manso-Simon, Y.; Batteas, J.; Marquez, M.; Ratner, M.;
Mujica, V. J. Phys. Chem. B 2004, 108, 18414-18420.
(21) Mujica, V.; Ratner, M. A.; Nitzan, A. Chem. Phys. 2002, 281, 147-150.
(22) Mujica, V.; Nitzan, A.; Datta, S.; Ratner, M. A.; Kubiak, C. P. J. Phys.
Chem. B 2003, 107, 91-95.
9
8310
J. AM. CHEM. SOC. 2007, 129, 8310-8319
10.1021/ja068729g CCC: $37.00 © 2007 American Chemical Society

Electron Tunneling Spectroscopy of a New Monolayer
A R T I C L E S
theory,23 using Green’s functions for quantum transport phe-
nomena and basis sets obtained by density-functional-theory or
Hartree-Fock methods for the molecule and various ap-
proximate models for the electrode’s density of states. This com-
putes the coupling of the molecule with the electrodes and
estimates the electron transport rates. However, charge trapping,
monoradical ion formation, and even adiabatic electronic
transitions in the molecule may also have to be considered.
One can assume that the total current passing through a MOM
junction has three components: (1) the elastic (i.e., lossless)
current due to tunneling through space, (2) the elastic current
due to tunneling through molecular orbitals, and (3) the inelastic
current due to electron scattering within the molecule.
Tunneling through space may be the largest contribution,
especially when the Fermi energies are not in resonance with a
relevant molecular orbital.24
Resonant tunneling current is expected to occur when the
Fermi level of the electrode is in resonance with (i.e., has the
same energy as) an unoccupied or partially occupied molecular
orbital. This through-molecule current, also called orbital-
mediated tunneling (OMT), can be a large part of the overall
current.25,26
In general, inelastic processes involve electron transfer
between metal and an organic or inorganic molecule O with
the formation of a stable molecular anion O- or cation O+; this
inelastic process is accompanied by a Franck-Condon rear-
rangement of the molecular geometry. Under these conditions,
the molecular transition (O f O+ or O f O-) is termed
adiabatic, and is generally slower. In contrast, an elastic process
implies a resonant transfer, with no change in molecular
geometry, and is associated with a vertical transition (O f O+
or O f O-) with no vibrational relaxation; such an elastic
process is generally faster.27
Inelastic effects, due to vibronic excitation, generally bring
a small contribution to the overall current, and occur at low
biases.28 In addition, electronic transitions within the molecule,
vertical or adiabatic, excited by the tunneling electrons, can
occur at higher voltages. These also bring a contribution to the
overall current, as seen in some electron tunneling spectroscopy
experiments.26,29,30
In this study, I-V measurements show that a Langmuir-
Schaefer (LS) monolayer of 1 (Figure 1A) exhibits current
asymmetries, i.e., rectification, when sandwiched between two
Au electrodes at room temperature. I-V measurements and
electron tunneling studies were carried out at 4.2 K on the same
LS monolayer of 1, now sandwiched between Al and Pb
electrodes. Electron tunneling spectroscopy can probe molecular
electronic states.29,30 At low biases, molecular vibrations are
expected and are observed as peaks in the second derivative of
the current with respect to voltage (d2I/dV2). At higher applied
voltages, molecular energy states were also observed and were
more intense for forward than for reverse bias. This effect
correlates well with the current asymmetries observed in the
I-V measurements. At much higher applied voltages, the
rectification reverses direction. A thorough analysis of the
monolayer orientation, vibration intensity changes for forward
and reverse bias, and other studies relevant to the electron
transport in MOM junction are also reported.
Experimental Details
The synthesis of 1 is documented in the Supporting Information
(Figures A and B).
Ultraviolet-visible (UV-vis) spectra were acquired with a Cary
50 single-beam spectrometer, using a quartz cuvette of 1 cm path length.
Fluorescence spectra were acquired with a Yvon-Jobin Fluoromax 2
spectrometer, using a quartz cuvette (1 cm path length) and a sample
concentration of 2.5 × 10-6 mol/L.
Near-infrared and infrared spectra were acquired with a Bruker IFS-
88 FTIR spectrometer, equipped with globar and tungsten sources and
cryocooled detectors: mercury cadmium telluride (MCT) for the mid-
infrared range and InSb for the near-infrared range. For each spectrum,
a 1000-scan interferogram was collected in a single-beam mode with
a 4 cm-1 resolution. The incident angle of the p-polarized light was
set to 87° relative to the substrate normal.
Pockels-Langmuir (PL) monolayers at the air-water interface and
LS and Langmuir-Blodgett (LB) transfers onto solid substrates were
carried out with a NIMA 622 computer-controlled film balance in a
room with HEPA-filtered air. The water subphase (18 MΩ cm) was
obtained from a Barnstead Easypure LF system.
XPS spectra were obtained with a Physical Electronics Apex XPS
system equipped with an Omicron EA125 hemispherical analyzer; the
sample analysis chamber had a base pressure of 7 × 10-9 Torr, and
Mg KR (1253.6 eV) radiation was used as the excitation source. The
scan parameters were set for a 25 eV pass energy and a 0.02 eV step
scan; six scans were collected and averaged for each high-resolution
spectrum.
The water contact angles of the LS films were determined with a
Rame´-Hart contact-angle goniometer. The contact angle was read in
one side of a drop on five different regions of the monolayer.
Variable-angle spectroscopic ellipsometry measurements were per-
formed on a J. A. Woolam spectroscopic ellipsometer in the wavelength
range 300-850 nm at 65°, 70°, and 75° angles of incidence.
The “Au|LS monolayer|Au” junctions used for electrical measure-
ments were made by thermal evaporation of 5 nm Cr, then 115 nm Au
onto a superpolished glass substrate (forming a film 120 nm thick),
followed by LS monolayer deposition and evaporation of a second
electrical contact by cold-gold technique5 in an Edwards 306A
evaporator (base pressure ≈ 10-7 Torr) equipped with a cold finger
cooled externally with liquid N2, and several temperature monitors and
quartz film thickness monitors. During the deposition of the second
electrical contact, on top of the organic monolayer, the substrates were
placed on the cold finger, which was kept at a measured temperature
of 150 K. The substrates were facing away from the evaporation source,
to avoid organic monolayer decomposition due to radiation. Each glass
slide would bear about 50 separate Au top “pads” 40 nm thick and
0.28 mm2 in area. For the IETS measurements the “Al|LS monolayer|Pb”
junctions were prepared similarly, except that the second electrical
contact, Pb, was evaporated with the sample facing directly the
evaporation source; during evaporation the substrate was also cooled
to 170 K, to avoid the organic monolayer decomposition due to hot Pb
atoms. During exposure of the Al substrates to moderate vacuum in
the evaporator, and then to atmospheric air and water vapor, an oxide
film of some thickness forms naturally atop the Al electrode. Each
“Al|LS monolayer|Pb” junction has a working area of 0.5 mm × 2
mm. The Al thickness was ∼150 nm; the Pb thickness was ∼600 nm.
Conventional I-V curves were measured with a Keithley 226
current-voltage source and monitor. The rectification ratio for any I-V
curve is defined by RR ) -I(Vmax)/I(-Vmax), where Vmax is the
maximum bias used to measure the I-V curve.
(23) Landauer, R. IBM J. Res. DeV. 1957, 1, 223-231.
(24) Simmons, J. G. J. Appl. Phys. 1963, 34, 1793-1803.
(25) Mazur, U.; Hipps, K. W. J. Phys. Chem. 1995, 99, 6684-6688.
(26) Mazur, U.; Hipps, K. W. J. Phys. Chem. B 1999, 103, 9721-9727.
(27) Zhou, X.-L., Zhu, X.-Y.; White, J. M. Surf. Sci. Rep. 1991, 13, 73-220.
(28) Lambe, J.; Jaklevic, R. C. Phys. ReV. 1968, 165, 821-833.
(29) Hipps, K. W.; Mazur, U. J. Am. Chem. Soc. 1987, 109, 3861-3865.
(30) Lu¨th, H.; Roll, U.; Ewert, S. Phys. ReV. B 1978, 18, 4241-4249.
9
J. AM. CHEM. SOC. VOL. 129, NO. 26, 2007 8311

A R T I C L E S
Honciuc et al.
Figure 1. (A) Molecular structure of fullerene-bis[ethylthio-tetrakis(3,4-dibutyl-2-thiophene-5-ethenyl)-5-bromo-3,4-dibutyl-2-thiophene] malonate, molecule
1 in fully extended conformation; length L ) long axis ≈ 7.0 nm, height H ≈ 1.5 nm, width W ) short axis ≈ 1.0 nm. The two electron donor regions
(thienylenevinylenes) are marked “D moiety”, and the electron acceptor region (fullerene) is marked “A moiety”. (B) Plane projection of the extended
conformation of 1, viewed from the fullerene side, showing the 1.0 nm thickness of the fullerene moiety and the minimum van der Waals thickness (0.35
nm) of the donor end. (C) Plane projection of a hypothetical maximally folded conformation, with thiophenes folded into a “U”: height H ≈ 4.0 nm, length
L ≈ 1.0 nm, width W ≈ 1.0 nm. (D) Closest packing of extended conformation B, with donor moieties interlocking and placed above fullerene regions of
adjacent molecules: estimated cross-sectional area ≈ 1.0 nm2. (E) Side view of “Au|LS monolayer of 1|Au” junction; the polarity shown on the electrodes
is for V > 0. The left arrow indicates the direction of electron flow at V > 0 for |V| e2 V, while the right arrow shows the direction of favored electron flow
for V < -2.5 V. (F) Side view of “Pd|LS monolayer of 1|Al” junction; the polarity of the electrodes is for V > 0. The arrow indicates the direction of
electron flow at V > 0.
IETS spectra were acquired at 4.2 K by second harmonic detection
in an immersion Dewar (BOC Edwards 60 L). The home-built IETS
spectrometer consisted of a dc voltage source, a lock-in amplifier (LIA),
a digital multimeter, a custom-made adder circuit, an ac source, and a
PC microcomputer (Windows 98) controlled by a program using the
Labview (National Instruments, Austin, TX, version 4.1) equipment
interface and data acquisition system (see Supporting Information,
Figure C). External sources of noise were minimized as well as possible.
The dc bias scan range was limited to below (1 V. Above this bias,
the noise matched or exceeded any possible signal. Signals were
obtained using a source modulation frequency of ω/2π ≈ 150-200
Hz and second harmonic detection of the (d2I/dV2) signal at 2ω. The
spectrometer performance was verified with benzoic acid, whose
measured spectra were in excellent agreement with the literature results
(see Supporting Information, Figures D and E). The IETS vibrational
peaks have an overall line width31 (full width at half-maximum) fwhm
) [(NLW)2 + (5.4kBT)2 + (1.7Vacmod)2]1/2 ) 6.882 mV ) 55.5 cm-1
for a natural vibration line-width NLW ) 1 mV, T ) 4.2 K, kB
)
(31) Wang, W.; Lee, T.; Reed, M. A. Rep. Prog. Phys. 2005, 68, 523-544.
9
8312 J. AM. CHEM. SOC. VOL. 129, NO. 26, 2007

Electron Tunneling Spectroscopy of a New Monolayer
A R T I C L E S
Figure 3. RAIRS spectrum of an LS monolayer of 1 deposited on Al.
Figure 2. FTIR spectrum of bulk 1 in a KBr pellet.
gaseous, up to an area of ∼250 Å2. Further compression brings
the molecules closer into a quasi-2-dimensional liquid phase,
which persists up to ∼200 Å2. The monolayer is fully
Boltzmann constant, and an applied ac modulation voltage Vacmod ) 4
mV. The dependence of fwhm on Vacmod was verified for the “Al|benzoic
acid|Pb” junctions.31 The energy conversion factor is 1 meV ≈ 8.065
compressed into a 2D-quasi solid phase above ∼10 mN m-1
.
cm-1
Results
Optical Spectroscopy of 1. The absorption spectrum of a
.
The area per molecule at zero film pressure is 270 Å2; at transfer,
it is 113 Å2. The solid-like PL monolayer was transferred onto
solid substrates via the LS horizontal technique at a pressure
of ∼20 mN m-2. The LS transfer ratios were high (110-120%),
probably because the molecules also deposited on the rougher
substrate edges. LS monolayers of 1 were deposited onto Au,
Al, and Si substrates.
solution of 1 in benzene shows a maximum at 529 nm, while
the fluorescence emission spectral maximum occurs at 622 nm
(Supporting Information, Figure F). The vertical band gap for
molecule 1 determined from the UV-vis spectrum absorption
edge is ∼2.0 eV.
Attempts to transfer the monolayer from the water by the
LB or vertical transfer method gave only low transfer ratios,
below 70%.
The IETS spectrum discussed below requires that optical
absorbance be measured also in the near-infrared (NIR) region.
No optically allowed electronic transition was observed in the
region between 650 and 3400 nm (Supporting Information,
Figure G). The most intense IR absorption bands are due to the
Molecular Orientation and Thickness Measurements.
Molecular orientation studies carried out with XPS-angle
resolved and grazing-angle reflection-absorption infrared spec-
troscopy (RAIRS). The RAIRS spectrum of the LS monolayers
deposited on the Al surface (Figure 3) shows important changes
in absorbance, when compared with Figure 2, due to surface
dipole selection rules. The aliphatic C-H vibrations at 2965
cm-1 can be taken as intensity reference bands, because they
are prone to conformational freedom on the surface. The vinyl
dC-H stretches present in the bulk at 3015 cm-1 (Figure 2)
are absent on the Al surface (Figure 3): these bonds are parallel
to the surface. The -CdC- and the CdO stretches of the ester
at 1646 and 1742 cm-1 are greatly enhanced, compared to bulk,
indicating that their bond dipoles are oriented perpendicular to
the metal surface. The readily identifiable intensities of the
various CdC modes of the thienylene rings and the vinyl group
in the 1457-1646 cm-1 region are also enhanced; the tetrakis-
(thienylvinylidene) moiety appears to have its long axis oriented
parallel with the metal surface, as depicted in Figure 1A. The
exact conformation of the two tetrakis(thienylvinylene) chains
is difficult to determine, because conformational movements
around the CdC bond of the vinylidene and isomerization of
C-C bond between vinylidene and thienylene ring can cause
disorder. Also, the XPS angle-resolved data (see Supporting
Information, Figures I and J) show that the outer surface of the
LS monolayer on Al is both C-rich and S-rich, indicating that
the S atoms are preferentially on the outmost surface of the LS
monolayer. The O atoms lie somewhat below the surface (nega-
tive slope of plot of the natural logarithm of the intensity versus
C-H asymmetric stretch at 2927 cm-1
.
The bulk FTIR spectrum of 1 in KBr is presented in Figure
2. The absorption band at 923 cm-1 corresponds to the vinylene
CH2 in-phase, out-of-plane wag.32,33 The very weak peaks at
1540, 1593, and 1646 cm-1 may be vinylene CdC stretch
modes. The -CH3 umbrella bending mode at 1373 cm-1 and
the n-butyl group -CH2 scissors mode33 at 1457 cm-1 are
intense. The vinyl CH2 stretch appears at 3015 cm-1, the
asymmetric sp3 CH2 stretch is at 2954 cm-1, the asymmetric
sp2 CH2 stretch is at 2921 cm-1, and the sp2 symmetric CH2
stretch appears at 2854 cm-1. The ester -CdO bond is a
prominent peak at 1746 cm-1. The ester C-O-C stretches are
at 1019 and 1101 cm-1. One Td C60 vibration34 is clearly seen
at 1181 cm-1. The other Td C60 vibration (expected34 between
1429 and 1432 cm-1) and the thiophene ring CdC stretch
(expected33 at 1436 cm-1) are here in an unresolved shoulder
of the very intense n-butyl -CH2 scissor peak at 1457 cm-1
.
Monolayer Deposition. The PL isotherm (Supporting Infor-
mation, Figure H) of molecule 1 shows three phases with the
monolayer compression at the air-water interface. The first
phase corresponds to the case where the molecules are dilute,
(32) Furukawa, Y.; Sakamoto, A.; Tasumi, M. J. Phys. Chem. 1989, 93, 5354-
5356.
(33) Bolognesi, A.; Catellani, M.; Musco, A.; Pontellini, R. Synth. Met. 1993,
55-57, 1255-1259.
(34) Chase, B.; Herron, N.; Holler, E. J. Phys. Chem. 1992, 96, 4262-4266.
9
J. AM. CHEM. SOC. VOL. 129, NO. 26, 2007 8313

A R T I C L E S
Honciuc et al.
Table 1. Measured Monolayer Thickness (by XPS and
Ellipsometry) and the Contact Angle of a Drop of Water over the
Monolayer
substrate
thickness XPS (nm)
thickness ellipsometry (nm)
water contact angle (deg)
Al
Si
Au
2.3 ( 0.8
93.8
92.5
92.5
1.89 ( 0.1
1.83 ( 0.1
the cosecant of the angle Θ). On the basis of the slope, ∼0.17,
and the inelastic mean free path within an organic monolayer,
which is typically between 22 and 27 Å for an electron kinetic
energy of 720 eV, one can estimate that O atoms are ap-
proximately 0.6 nm below the outermost LS monolayer surface.
On the basis of this spectroscopic evidence, the molecular
orientation is such that the C60 moieties are closest to the surface
of the metal surface, while the carbonyl groups of the malonate
ester are perpendicular to the surface, and the tetrakis(thienylene
vinylidene)s are placed so that their long axes stretch sideways
from the fullerene, possibly parallel to, or at small angles to,
the metal surface. The measured organic monolayer thickness
data obtained by XPS depth profiling and by spectroscopic
ellipsometry are presented in Table 1, together with the contact
angle of a water droplet atop the monolayer. The organic
monolayer was modeled as two Cauchy layers with refractive
indices 2 and 1.5.
Figure 4. I-V curve for “Au|monolayer of 1|Au” junction (no. 18) at
room temperature and for a smaller bias range (|V| e 1.5 V). The turn-on
bias for rectification is V ≈ 0.8 V. Inset: rectification ratios RR ) I(Vmax)/
I(-Vmax) for all I-V measurements at room temperature for Vmax e 3 V.
Monolayer Structure and Orientation within the MOM
Junctions. The structure of the PL and LS monolayer can be
deduced by combining the pressure-area isotherm (Supporting
Information, Figure H) with the XPS and FTIR results. The
molecular structure, shown in Figure 1A, has been optimized
by AM1 calculations (HyperChem), which yield a linear donor
structure atop the C60 acceptor, with approximate dimensions
L ≈ 7.0 nm, W ≈ 1.0 nm (corresponding to the diameter of C60
) 1.0 nm),35 and H ≈ 1.5 nm. The van der Waals thickness of
the oligothienylenevinylenes, if planar, is 0.35 nm (Figure 1B).
The C60 moiety is the more hydrophobic group. The LS
monolayer thickness is 1.9 nm (Table 1), or slightly less than
two C60 diameters; this means that the two donor arms cannot
fold and extend vertically above the C60 acceptor (Figure 1C),
because this would yield a film thickness of 4.0 nm. The XPS
data also indicate that the O and S atoms are at 0.6 and 0 nm,
respectively, from the outermost monolayer surface. The area
per molecule in the LS monolayer at the transfer pressure
estimated from the Langmuir isotherm (Figure H) is At ) 1.13
nm2, which is 11% larger than the area of one C60 molecule. If
molecules 1 are accosted side by side, with no overlap, the area
per molecule is estimated to be ≈2.5 nm2, i.e., much larger than
the area at the transfer pressure. However, if the donor arms lie
somewhat above the C60 eicosahedra of neighboring molecules,
and exhibit the intermolecular π-π cofacial packing that is
observed in some 3,4-n-butyl substituted oligothiophene crys-
tals,36,37 then the interlocked structure of Figure 1D yields the
desired area ≈1 nm2, and a minimum film thickness of 1.5 nm.
With some bending of the thienylenevinylenes, this thickness
can easily become 1.9 nm. This interlocked structure explains
all the measured data for the PL and LS monolayers, and
Figure 5. I-V curve for “Au|monolayer of 1|Au” junction (no. 9H2) at
room temperature and larger bias range. The turn-on bias for rectification
is V ≈ -2.5 V. Inset: Reverse rectification ratios RRR ) I(-Vmax)/I(Vmax
)
for I-V measurements at room temperature for Vmax g 3 V.
explains why LB transfer is impossible: the film is too rigid,
and can be transferred only by quasihorizontal LS transfer.
Figure 1E,F shows how an LS monolayer of 1 can be
arranged between Au electrodes (Figure 1E) and between Pb
and Al electrodes (with the concomitant oxide coverings) (Figure
1F), and indicates, as explained below, the direction of electron
flow through the monolayer at forward bias (0 < V < 2.0 V)
or at large reverse bias (V < -2.5 V).
I-V Measurements. The LS monolayers of 1 on Au were
used for electrical I-V measurements at room temperature in
“Au|LS monolayer of 1|Au” junctions, while the “Al|LS
monolayer of 1|Pb” junctions were used for I-V measurements
at 4.2 K.
The LS monolayers of 1 sandwiched between two Au
electrodes exhibit current rectification at room temperature
(Figures 4 and 5). But there are two regimes. (i) For Vmax e 2
V, the higher currents are at positive bias (V > 0) (Figure 4)
with rectification ratios (RR ≡ -I(Vmax)/I(-Vmax) between 3
and 5 (forward bias corresponds to the top electrode (Au) being
biased positively, and the bottom Au electrode biased nega-
tively). (ii) For Vmax g 2 V, the higher currents are at negative
(35) Yang, S. H.; Pettiette, C. L.; Conceic¸ao, J.; Chesnovsky, O.; Smalley, R.
E. Chem. Phys. Lett. 1987, 139, 233-238.
(36) Huang, W.; Masuda, G.; Maeda, S.; Tanaka, H.; Ogawa, T. Chem. Eur. J.
2006, 12, 607-619.
(37) Pappenfus, T.; Raff, J. D.; Hukkanen, E. J.; Burney, J. R.; Casado, J.; Drew,
S. M.; Miller, L. L.; Mann, K. R. J. Org. Chem. 2002, 67, 6015-6024.
9
8314 J. AM. CHEM. SOC. VOL. 129, NO. 26, 2007

Electron Tunneling Spectroscopy of a New Monolayer
A R T I C L E S
Figure 7. I-V curve of an “Al|LS monolayer of 1|Pb” junction (no. 13)
at 4.2 K (RR ) 5.33). The turn-on voltage is about 0.6 V. This is the same
junction as measured by IETS in Figure 9. Inset: Histogram of the
rectification ratios of “Al|LS monolayer of 1|Pb” junctions at 4.2 K
Figure 6. (A) Fowler-Nordheim plot of part of the I-V data for
“Au|monolayer of 1|Au” junction (no. 1H3.1) at room temperature: poor
least-squares linear fit for data measured for increasing positive bias, from
V ) 1 V to V ) 3.5 V; very good least-squares linear fit (R) 98.7%) for
data measured thereafter at decreasing bias, from V ) 3.5 V to V ) 1.0 V.
(B) Same plot, but for negative bias: good fit for data from V ) -3.5 V
to V ) -1.0 V.
bias (V < 0). At bias ranges above (2.5 V, the rectification
changes sign, with higher currents for V < -2.5 V (Figure 5),
as observed recently for a chemically very similar fullerene
derivative:9 the currents are comparatively higher, and the
reverse rectification ratios are larger in this higher-bias regime.
A summary of all I-V data with Au electrodes is given in
Table A of the Supporting Information, which also has two more
I-V curves (Supporting Information, Figures K and L).
When the applied biases exceed the tunneling barrier height
ELUMO - EF(Au), then cold-electron emission can also occur from
the rougher electrode (e.g., the top electrode). This emission
has been described by the Fowler-Nordheim (FN)24,31,38,39
equation JV-2 ) (144e3/529πhBd2)exp(-231/2πm1/2B3/2s/6ehV),
where J is the current density, V is the voltage (V), B is the
tunneling barrier relative to the electrode Fermi level EF(Au), e
is the electronic charge, m is the mass of the electron, h is
Planck’s constant, and d is the gap between electrodes (here d
) 19 Å). As seen in Figure 6A, the FN equation fits rather
well for relatively high potentials (positive and negative), upon
return from the extrema (Vmax or -Vmax) toward lower voltages,
but fits poorly when these extrema are first approached from
zero bias. In Figure 6 the absolute value of the current density
J is 1.0 A m-2 at V ) |3.5 V|. The slopes of the least-squares
straight line fits are significantly different for positive and
negative bias: the currents are larger when the electrons travel
1.9 nm from the “cold-gold” top Au electrode, which has the
rougher surface, to the smoother bottom Au electrode. Is this
asymmetry in surface roughness enough to explain the asym-
metry in currents at room temperature seen in Figure 5?
For comparisons with the IETS reported below, the I-V curve
was also measured for an “Al|LS monolayer of 1|Pb” junction
at 4.2 K (Figure 7). The I-V curve is asymmetric, RR ∼ 5.
This correlation between the rectification ratio and the absolute
intensity ratio of the broad bands, at negative and positive biases,
holds well for most of the MOM junctions analyzed. However,
Figure 8. The IETS vibrational region spectrum (d2I/d2V) of an “Al|LS
monolayer of 1|Pb” junction (no. 9). The spectrum was acquired at 4.2 K
(ac excitation signal Vacmod ) 6.4 mV, frequency ν ) 170 Hz).
a few junctions did not show broad bands, although they
exhibited small current asymmetries in the I-V curves (RR <
2). For applied voltages below 1.0 V, the rectification ratios
are naturally slightly smaller, between 1 and 3. A histogram
(inset to Figure 7) of the I-V curves and the distribution of the
number of junctions analyzed with the RR of the LS monolayers
of 1 sandwiched between Al and Pb electrodes indicate that
the RRs converge to RR ≈ 2.
A possible FN regime for the “Al|LS monolayer of 1|Pb”
junction at 4.2 K cannot be discussed here, since the required
high-voltage regimes were not accessible in our IETS measure-
ments.
IETS Measurements. The IETS spectrum for an LS mono-
layer of 1 sandwiched between Al and superconducting Pb
electrodes is given in Figure 8 for the energy range 0.06-0.40
eV (500-3200 cm-1). All the molecular vibrations, either
Raman-active or FTIR-active, appear together in the spectrum,
since electric dipole selection rules do not apply for IETS. The
energy resolution for the IETS spectra, calculated as ∼55 cm-1
,
is inferior to that of FTIR or Raman spectroscopy (less than
∼8 cm-1). IETS peaks are seen at 572, 680, 783, 890, 1260,
1337, 1400, 1573, 1701, 1784, 2877, and 3031 cm-1. These
(38) Fowler, R. H.; Nordheim, L. Proc. R. Soc. London 1928, 119, 173.
(39) Simmons, J. G. J. Appl. Phys. 1963, 34, 2581-2590.
9
J. AM. CHEM. SOC. VOL. 129, NO. 26, 2007 8315

A R T I C L E S
Honciuc et al.
34
Table 2. Comparison of IETS, RAIRS, and Bulk FTIR Spectra of 1 with Raman Spectra of Poly(thienylenevinylene) (PTV)32,40 and of C60
FTIR Spectrum of Poly(3,4-dibutylthienyl-enevinylene) (PbuTV),33 IETS of Stearic Acid, C17H35COOH (C17Ac),41 and of n-Octanethiol,
C8H17SH (C8SH),42 and Interpretation of the IETS Spectrum of 1
,
FTIR
RAIRS
IETS
Raman
PTV32
Raman
PTV40
FTIR
PbuTV33
Raman
IETS
C17Ac.41
IETS
C8SH42
interpretation of
34
bulk 1
mono 1
mono 1
C60
IETS of mono 1
572
680w
783w
557
654
741
569
772
Td of C60
R def
R def
700
797
649
929
733
778
832
834
843
882
902
863
923
926w
993
890
919
V CH bend
950
996
1019w
1040
1045
1032
1074
1113
1140
1045
1285
1073
1274
1101w
1181w
1229
1099w
1228
1100
1250
1169
1212
1239
1275
1287
1288
1350
1354
1410
1412
1488
1260
V in CH bend
1312
1373m
1457
1373m
1459
1337w
1400
1367
1448
1409
1584
1436
1464
1425
R CHstr/C60
1468
1481
1574
1468
1601
1500
2879
1540w
1593w
1646w
1746
1540
1573
1586
V CdCstr/C60?
1646
1742
1701
1784
2877
ester C-O str
ester CdO str
asym CH2 str
2854
2921
2954
2859
2955
2917
3013
as V CH str
peaks can be identified by comparison with literature32-34,40-42
(Table 2) as follows: a weak vinylene C-H stretch at 3031
cm-1, an intense symmetric CH3 stretch at 2877 cm-1; an ester
CdO stretch at 1784 cm-1; an ester C-O stretch at 1701 cm-1
,
shifted to somewhat higher energies than in FTIR; a combination
of a Td vibration of C60 with a vinyl CdC stretch at 1573 cm-1
a thiophene ring C-H stretch at 1400 cm-1; a vinyl C-H bend
at 1260 cm-1; a nonplanar vinyl C-H bend at 890 cm-1
thiophene ring deformations at 783 and 680 cm-1; and a Td
vibration of the C60 moiety at 572 cm-1
;
;
.
The Broad Bands. At larger biases (Figure 9), antisymmetric
broad bands appear at ∼ +0.66 V and also at ∼ -0.60 V. The
ratio between the intensity of these bands, at positive and
negative bias, measured from zero (Figure 9), is ∼1.7. This
broad band is more intense at forward bias, i.e., in same direction
as the preferred electron flow observed in the I-V curves (Figure
7). The asymmetry in the broad bands is preserved in a
normalized tunneling intensity (NTI) spectrum representation
(inset of Figure 9); the sloping background is removed by
plotting (d2I/dV2)/(dI/dV), instead of (d2I/dV2), versus V.29
Figure 9. Wider-range IETS scan of an “Al|LS monolayer of 1|Pb” junction
(no. 13), showing the (a) ZBA, (b) dominant CH2 peaks, (c) electronic
artifacts marked “X”, and (d) broad OMT feature at ∼0.65 V: RR(V )
0.65 V) ) 1.8. This is the same junction as measured in Figure 7. The
inset shows the normalized tunneling intensity (NTI) spectrum, showing
that the asymmetries of the OMT bands are preserved after background
removal.
polarities: The intensities are larger for Al electrode biased
positively. This could be due to the molecular orientation: the
CH groups are closer to the Pb electrode and couple better with
the Pb electrodes. The organic monolayer is separated from the
bare Al electrode by Al2O3. Such asymmetry has been discussed
theoretically.43
The vibronic features of the IETS spectra acquired for the
positive and negative biases are antisymmetric. The C-H bond
vibration peaks appear to have different intensities at opposite
(40) Mevellec, J. V.; Buisson, J. P.; Lefrant, S.; Eckhardt, H. Synth. Met. 1991,
41-43, 283-286.
(41) Brousseau, J.-L.; Vidon, S.; Leblanc, R. M. J. Chem. Phys. 1998, 108,
7391-7396.
(42) Wang, W.; Lee, T.; Kretschmar, I.; Reed, M. A. Nano Lett. 2004, 4, 643-
(43) Galperin, M.; Nitzan, A.; Ratner, M. A.; Stewart, D. R. J. Phys. Chem. B
2005, 109, 8519-8522.
648.
9
8316 J. AM. CHEM. SOC. VOL. 129, NO. 26, 2007

Electron Tunneling Spectroscopy of a New Monolayer
A R T I C L E S
At near-zero bias, intense and narrow bands appear: These
have been traditionally called zero-bias anomalies (ZBA), and
are presumably due to Pb phonons within the superconducting
Pb electrode.44
of d2I/dV2, as in IETS. Therefore, for the particular molecular
junction investigated here the broad bands observed can be
understood in terms of current flow enhancement, because at
some bias the Fermi level of the electrode and a relevant
molecular orbital are brought to the same energy. For molecule
1, the probed relevant molecular orbital, i.e., the LUMO,
mediates the electron tunneling at both positive and negative
bias. At positive bias the potential, at which this broad band
appears (0.6 V), is roughly twice the EF(Al) - LUMO gap; at
negative bias the band appears at -0.66 V, or twice the EF(Pb)
- LUMO gap. On the basis of the above considerations, the
LUMO energy is placed at roughly -3.9 eV with respect to
vacuum. In the above calculations, the electrode work-functions
were 4.17 eV for Al and 4.2 eV for Pb, as determined from
photoelectron measurements.49 The difference in IETS intensities
at +0.60 and -0.66 V, and the difference in currents in the
I-V curves at these two potentials, can be explained by the
greater proximity of the LUMO to the Al electrode than to the
Pb electrode (neglecting the defective Al oxide, for IETS) or
to the bottom Au electrode (for the I-V curves), which makes
for a more efficient coupling at forward bias (V > 0) than at
reverse bias. The width of the two OMT bands is due to either
(a) the polycrystallinity of the metal electrodes, whose Pb and
Al crystallites have random orientation, or (b) the vibronic
envelope that widens any electronic transition in the molecule.
For explanation (a), the broadening would be due to a sample-
average range of crystallite orientations with a concomitant range
of work-functions. Indeed, the Al work-function values are 4.20
eV for the (100) face, 4.06 eV for the (110) face, and 4.26 eV
for the (111) face, is ∼4.17 eV, while the value for polycrys-
talline Pb is 4.25 eV.49 For explanation (b), Figure F suggests
a 1 eV width for an allowed optical transition.
On a different junction (Supporting Information, Figure M),
RR ) (signal at 0.65 V/signal at -0.65 V) is smaller than in
Figure 9, and the ZBA signals are also weaker. Two other
junctions have essentially identical features to Figure 9 (Sup-
porting Information, Figures N and O).
In Figure 9, the OMT bands appear at the same “turn-on”
voltage obtained from the interpolation of the I-V curve with
the voltage, Figure 4, i.e., at ca. 0.7 V at forward bias.
A series of oscillations in the inelastic tunneling spectra of
“Al|Al2O3|Pb” junctions have been studied in detail, and were
dubbed the “quantum size effect” (QSE).45-47 The QSE oscil-
lations were observed between 0.4 and 1.4 V, with a center at
0.85 V,45 and are due to particle-in-a-box standing-wave states
in the superconducting Pb electrode;45 they need a well-formed
aluminum oxide barrier, depend explicitly on the Pb electrode
thickness, and are absent for V < 0.46 A plot of the QSE
oscillation periods is linear with the reciprocal of the Pb layer
thickness.45 For the Pb layer thickness of 600 nm used here,
the period of these oscillations should be 0.037 V.45 Since no
such period is seen in Figure 9, or in Supporting Information
Figures M, N, or O, and since a weaker broad band is visible
for V < 0, therefore46 the broad bands of the present study are
not due to QSE oscillations.
The broad bands observed in this study could be ascribed to
either (1) inelastic scattering, due to an electronic transition in
the molecule, or (2) elastic scattering, i.e., as a consequence of
the enhanced electron flow due to resonance.
In an unrelated organometallic system, low-energy (∼5000
cm-1) dipole-allowed and spin-allowed optical electronic transi-
tions were observed, both optically and in IETS. The latter,
elastic peak was given the name “orbital-mediated tunneling”
(OMT).29 Here, spin-allowed and dipole-allowed inelastic
processes involving an electronic transition are ruled out,
because around 5200 cm-1 (0.65 eV) the NIR spectrum of 1
(Supporting Information, Figure G) shows no electronic transi-
tion; also, the chemical structure of 1 makes an optical absorp-
tion unlikely at such low energies. The addition to, or subtraction
from, a molecular orbital could be inelastic, but would be
observable in electrochemical (not optical) experiments.
If the above interpretation of the broad band is correct, then
one can probe with IETS the molecule’s affinity level (electron
affinity) and the creation of a molecular anion. OMT bands have
been previously observed in IETS; the process of “molecular
excitation to an affinity level” and determination of the “vertical
attachment energies” were previously suggested by Hipps and
Mazur.25,50
The LS monolayers of 1 are structured, and the molecules
are oriented asymmetrically on the surface: Thus, the current
asymmetry observed in IETS may be mostly due to the different
tunneling efficiencies from the “Pb|LS monolayer of 1” side
versus the “Al|LS monolayer of 1” side of the MOM junction.
The second, elastic hypothesis may appear unusual, since the
name of the method is IETS. However, the electron tunneling
measurements with a lock-in amplifier measure the increase in
the total current, rather than an abrupt change in the I-V slope
(“kink”) at a particular voltage, caused by a scattering process
and opening of a second tunneling channel.48 A resonant
tunneling process, in which an empty molecular orbital is
brought into resonance, at a particular voltage, with the Fermi
energy of a metal would also produce a change in the I-V slope,
and thus will be observed in the second harmonic measurement
Among all the MOM junctions of 1 studied so far, the ratio
between the broad-band intensities, in the d2I/dV2 versus V
spectrum, for the forward versus the reverse bias, is found to
be closely related to the rectification ratio observed in the I-V
curves. The higher band at forward bias is always more intense.
The forward bias occurs in the same directions for I-V (|V| e
2 V, Au electrodes) and IETS (Pb and Al electrodes). The
fullerene moieties are closest to the negative electrode, and
the thienylene-vinylene moieties are closest to the positive
electrode. Previous studies51 of current asymmetries in Al-
AlOx-Pb junctions, with various randomly oriented organic
(44) Wolf, E. L. Principles of Electron Tunneling Spectroscopy; Oxford
University Press: Oxford, U.K., 1985; p 393.
(45) Jaklevic, R. C.; Lambe, J.; Kirtley, J.; Hansma, P. K. Phys. ReV. B 1977,
(49) CRC Handbook of Chemistry and Physics, 82nd ed.; Lide, D. R., Ed.; CRC
Press, Boca Raton, FL, 2001-2002; p 12-130.
15, 4103-4104.
(46) Jaklevic, R. C.; Lambe, J. Phys. ReV. B 1975, 12, 4146-4160.
(47) Kipps, K. W.; Susla, B. P.; Dunkle, E. J. Phys. Chem. 1986, 90, 3898-
3900.
(50) Hipps, K. W.; Mazur, U. J. Phys. Chem. 1994, 98, 5824-5829.
(51) Muldoon, M. F.; Dragoset, R. A.; Coleman, R. V. Phys. ReV. 1979, B20,
416-429.
(48) Hansma, P. K. Phys. Rep. 1977, 30, 145-206.
(52) Honciuc, A. Ph.D. Dissertation, University of Alabama, August 2006.
9
J. AM. CHEM. SOC. VOL. 129, NO. 26, 2007 8317

A R T I C L E S
Honciuc et al.
Figure 10. Proposed mechanisms for the two rectification modes of D-σ-A molecule 1. The (bottom) metal electrode M1 is grounded throughout. (A) At
zero bias, the HOMO, mostly localized on the bis-tetrakis(thienylenevinylene)s, is closer to the bottom electrode, and the LUMO, mostly localized on the
fullerene, is closer to the top electrode M2. (B) At V g 0.7 V, the LUMO comes into resonance with the Fermi level of M1, and resonant tunneling occurs
through the orbital to the second electrode M2. (C) At V ) -2.5 V the LUMO comes in resonance with M2, and the HOMO is in resonance with M1: An
electron moves from M2 to LUMO (step 1), and an electron from the HOMO moves to M1 (step 2, presumably in synchrony with step 1). This creates the
excited state D+-σ-A-, which relaxes somehow to the ground state D0-σ-A0.
molecules interspersed between the alumina and Pb layers, show
that the I-V characteristic of a junction is strongly dependent
on the adsorbate; the corresponding broad bands observed in
IETS spectrum were attributed to “low-lying organic barrier
heights”, i.e., to EF(metal) - LUMO.51
measurement). In the following simplified discussion of electron
transfer through an organic monolayer, it is assumed that the
potential drop between the two metal surfaces occurs exclu-
sively, but not linearly, across the monolayer.9 It is further
assumed that the oxide that formed on the Al electrode upon
exposure to air is sufficiently thin and defective that an energy
barrier across the oxide can be ignored.
Discussion
We now have to place the HOMO and LUMO of the
monolayer, relative to the metal work functions. The optical
spectrum (Figure F) leads to a HOMO-LUMO gap of 2.2 eV.
The onset of asymmetric conductivity suggests that the LUMO
should be placed at -3.9 eV, even though the experimental gas-
phase electron affinity of C60 is 2.8 eV.35 Placing the HOMO
2.2 eV below the LUMO, at -6.1 eV, is also suggested by our
I-V data, even though the gas-phase ionization potential of the
oligothienylenevinylenes is probably around 7.5 eV. To account
for the molecular orientation, we have pinned the LUMO and
the HOMO energies to the adjacent metallic electrodes M1 and
M2, respectively. As the bias on electrode M2 departs from zero,
the HOMO and LUMO energies must change with the applied
electric field within the 1.9 nm thick monolayer, as the Gibbs
chemical potential (or Fermi level) shifts in response to the
applied bias. This shift may not be linear with distance across
the monolayer.9
The asymmetry in Figure 10B suggests that the current at
moderate forward bias is larger than that at reverse bias, perhaps
because the fullerene moiety is physically closer to the bottom
Au electrode M1 than to the top cold-gold evaporated Au
electrode M2. In contrast, in Figure 10C, at high reverse bias
the favored electron transfer (M2 to LUMO, HOMO to M1) is
to orbitals placed farther from the appropriate electrode.
While the work-functions for Pb and Al are very close to
each other, there is a significant difference in the work-functions
IETS measurements have yielded not only (1) the vibrational
spectrum of an LS monolayer of 1, but also (2) a broad and
asymmetric signal due to the onset of molecular rectification
and resonant tunneling between the metal electrodes and the
molecule. This is the conclusive proof that the rectification we
have been measuring for several years is indeed due to through-
bond tunneling processes within the molecules.
The LS monolayer of 1 also shows two rectification regimes,
one at relatively lower bias, and a different one at higher bias.
(1) For Vmax e 2 V the enhanced currents are for V > 0 [turn-
on at 0.8 V for Au electrodes (Figure 4) and at 0.6 V for Pb
and Al electrodes (Figure 7)], and (2) for Vmax g 3 V and Au
electrodes, the higher currents occur are for V < 0 [turn-on at
-2.5 V] (Figure 5). The first rectification is most likely due to
electrons accessing only the unoccupied orbitals, i.e., the LUMO,
which should be localized on the fullerene moiety,9 while the
second rectification is probably due to two electron transfers,
one from the metal electrodes to the LUMO on the fullerene
acceptor moiety and the other from the HOMO on the
thienylene-vinylene electron donor moiety to the Au electrode.
This interpretation is schematically presented in Figure 10. The
metal electrode M1 is held at zero applied potential throughout.
The work function of Au was taken as 4.70 eV (in the literature
there is considerable variation of this value, from 4.70 to 5.40
eV, depending on the crystal face of Au and the method of
9
8318 J. AM. CHEM. SOC. VOL. 129, NO. 26, 2007

Electron Tunneling Spectroscopy of a New Monolayer
A R T I C L E S
Conclusion
of Au and Al (0.5-0.8 eV, depending on method and crystal
facet): It is thus a bit surprising that the “turn-on” voltages are
0.8 V (Figure 4) for Au, but 0.60 V for Al + Pb (Figure 7).
Previous experience5 suggests that this difference (0.20 V)
should be at least 0.5 V.
The rectification by a single molecule in an LS monolayer
has been confirmed by d2I/dV2 measurements at 4.2 K between
Al and Pb electrodes as due to electrons from the metal electrode
transferring resonantly to the available LUMO of the molecule.
This orbital-mediated tunneling is asymmetric and opposite in
sign with respect to junction polarities, and occurs at the same
voltage as the onset of enhanced current flow in direct-current
measurements. IETS vibrational spectra were also measured.
At much higher potentials, the current asymmetry is reversed:
At sufficiently high bias, both HOMO and LUMO can be
accessed from the metal electrodes.
The good fit of the FN equation in Figure 6A gives, from its
slope, a tunneling barrier height of ∼0.4 eV above the EF(Au)
,
thus placing LUMO at -4.3 eV; this estimate places the LUMO
0.4 eV lower than the estimates from IETS (∼ -3.9 eV) or
from the energy levels chosen in Figure 10 (∼ -3.9 eV). The
HOMO may also be involved in the electron transport across
the “Au|LS monolayer of 1|Au” junction; this is also suggested
by the strong hysteresis between the scans in opposite scan
directions, and the bad fits of the FN equation for the return
curves in Figure 6. Unlike resonant tunneling through a LUMO,
which is an empty orbital readily available to an incoming
electron, tunneling through an initially filled HOMO should
involve two consecutive steps: first, “half-emptying” the
HOMO, and next, “refilling” it.
Acknowledgment. We are grateful to Mr. Gihan Kwon for
assistance with XPS data collection, and to Professors Patrick
LeClair (University of Alabama), Gregory J. Szulczewski
(University of Alabama), Mark A. Reed (Yale University), and
Kerry W. Hipps (Washington State University) for helpful
discussions. The IETS equipment was acquired using residual
funds of Grant NSF-DMR-00-95215. A.G. was supported by
DOE-EPSCoR.
For the positive bias case, +1 to +3.5 V in Figure 6A, where
the HOMO is physically closer to the M2 electrode than to the
M1 electrode, as shown in Figure 10C, the electron transfer from
the LS monolayer of 1 to M2 (emptying) may be faster than
the transfer from M1 to LS monolayer of 1 (filling). That is,
the extraction of an electron from the HOMO may be somewhat
adiabatic (“charge trapping”). This would contribute to the
observed hysteresis in Figure 6A.
For the negative-bias regime, the FN fit, Figure 6B, gives a
barrier height of ∼0.7 eV above the EF(Au), and a LUMO at ∼
4.0 eV with respect to vacuum level, in better agreement with
previous estimates. Here, the HOMO contribution to electron
transport may be less, and less hysteresis is seen.
Barring further data analyses and experiments on other
systems and, most importantly, upon overcoming experimental
barriers in IETS, such as making low-resistance contacts to
“Au|monolayer|Au” junctions that can remain ohmic under He
temperatures, we consider the mechanism presented in Figure
10 to be a valid model of LS monolayer rectification.
Supporting Information Available: Synthesis and charac-
terization details, a schematic diagram of the IETS spectrom-
eter,52 IETS spectra of benzoic acid,52 the UV-vis and
fluorescence spectra of solutions of 1, the near-IR spectrum of
a film of 1, the PL isotherm of 1, the XPS intensities for a
monolayer of 1 as functions of the takeoff angle, a table of all
the I-V measurements of monolayers of 1 between Au
electrodes at room temperature, two I-V curves of a monolayer
of 1 between Au electrodes at room temperature, IETS spectra
of 1 between Al and Pb electrodes at 4.2 K for a narrower-
range scan (-0.6 to 0.6 V) and for wider-range scans (-0.7 to
0.7 V). This material is available free of charge via the Internet
at http://pubs.acs.org.
JA068729G
9
J. AM. CHEM. SOC. VOL. 129, NO. 26, 2007 8319