Self-decoupled porphyrin with a tripodal anchor for molecular-scale electroluminescence

Article abstract of DOI:10.1021/ja4048569

A self-decoupled porphyrin with a tripodal anchor has been synthesized and deposited on Au(111) using different wet-chemistry methods. Nanoscale electroluminescence from single porphyrin molecules or aggregates on Au(111) has been realized by tunneling electron excitation. The molecular origin of the luminescence is established by the vibrationally resolved fluorescence spectra observed. The rigid tripodal anchor not only acts as a decoupling spacer but also controls the orientation of the molecule. Intense molecular electroluminescence can be obtained from the emission enhancement provided by a good coupling between the molecular transition dipole and the axial nanocavity plasmon. The unipolar performance of the electroluminescence from the designed tripodal molecule suggests that the porphyrin molecule is likely to be excited by the injection of hot electrons, and then the excited state decays radiatively through Franck-Condon π*-π transitions. These results open up a new route to generating electrically driven nanoscale light sources.

Full text of DOI:10.1021/ja4048569

Article
pubs.acs.org/JACS
Self-Decoupled Porphyrin with a Tripodal Anchor for Molecular-Scale
Electroluminescence
San-E Zhu,†,‡ Yan-Min Kuang,† Feng Geng,† Jia-Zhe Zhu,† Cong-Zhou Wang,†,‡ Yun-Jie Yu,† Yang Luo,†
Yang Xiao,†,‡ Kai-Qing Liu,†,‡ Qiu-Shi Meng,† Li Zhang,† Song Jiang,† Yang Zhang,† Guan-Wu Wang,*,†,‡
Zhen-Chao Dong,*,† and J. G. Hou†
‡
†Hefei National Laboratory for Physical Sciences at the Microscale and CAS Key Laboratory of Soft Matter Chemistry and
Department of Chemistry, University of Science and Technology of China, Hefei, Anhui 230026, China
S
* Supporting Information
ABSTRACT: A self-decoupled porphyrin with a tripodal anchor has been synthesized and
deposited on Au(111) using different wet-chemistry methods. Nanoscale electroluminescence from
single porphyrin molecules or aggregates on Au(111) has been realized by tunneling electron
excitation. The molecular origin of the luminescence is established by the vibrationally resolved
fluorescence spectra observed. The rigid tripodal anchor not only acts as a decoupling spacer but
also controls the orientation of the molecule. Intense molecular electroluminescence can be
obtained from the emission enhancement provided by a good coupling between the molecular
transition dipole and the axial nanocavity plasmon. The unipolar performance of the
electroluminescence from the designed tripodal molecule suggests that the porphyrin molecule is
likely to be excited by the injection of hot electrons, and then the excited state decays radiatively
through Franck−Condon π*−π transitions. These results open up a new route to generating
electrically driven nanoscale light sources.
controlled orientations. We shall describe the design strategy of
a self-decoupled porphyrin molecule (abbreviated as 1) bearing
a rigid tripodal anchor. The challenges of its synthesis,
deposition on metals, and clear characterization by STML
will be addressed. Interestingly, different wet-chemistry
deposition methods have been found to affect the adsorption
configuration of molecules on the metal surface and resultant
electroluminescent properties. A new understanding of the
excitation mechanism of molecular electroluminescence in the
STM junction will also be discussed briefly.
INTRODUCTION
■
The highly localized tunneling current in a scanning tunneling
microscope (STM) can be used as a source of low-energy
electrons to locally excite photon emission from metals,
semiconductors, and organic molecules.1 In recent years,
STM-induced luminescence (STML) from organic molecules
on metals has gained much attention for its capability in helping
to develop nanoscale molecular optoelectronics and plas-
monics. However, when molecules are directly adsorbed on or
very close to metal surfaces, the intrinsic molecular fluorescence
will be quenched by the strong interaction between the
molecules and the substrate.2−9 Decoupling the molecule from
the substrate underneath is essential to the generation of
molecule-specific electroluminescence near metals. To date,
such decoupling has usually been realized by inserting an
additional spacer layer such as an oxide, halide, or molecular
multilayer in between.2−5,10,11 The idea of chemically
decoupling the emitter from the metal substrate has been
proposed using a double-decker molecule by Berndt et al.12,13
The molecule thus decoupled is found to modify the plasmonic
emission profile dramatically. Nevertheless, the plasmonic
nature in the observed emission suggests that the decoupling
strength may still not be sufficient to generate intrinsic
molecular fluorescence. Sufficient electronic decoupling and
favorable axial dipole orientations are both believed to be
important factors in the generation of molecule-specific
fluorescence.8,10,11,14,15 In this work, we explore the chemical
approach to achieving molecule-specific emission by building
the spacer and emitter together within a single molecule with
RESULTS AND DISCUSSION
■
Figure 1 illustrates schematically the design principle for the
self-decoupled porphyrin molecule (1) on Au(111). The
porphyrin moiety is selected as the emitter because it has
proven effective in producing molecular electroluminescence in
the STM junction.10 The rigid tripodal anchor is composed of a
stiff sp3-hybridized carbon tetrahedral framework and is
designed not only to bind the molecule on the Au surface
through the formation of three robust Au−S bonds but also to
serve as the decoupling spacer together with the linking unit of
an ethynyl group. The latter is again connected to the
peripheral phenyl ring of the porphyrin (Scheme 1). In this
way, the porphyrin core plane will be oriented along the tip axis
once the molecule stands up as designed and will thus allow for
optimal coupling between the molecular transition dipole16 and
Received: May 14, 2013
Published: September 25, 2013
© 2013 American Chemical Society
15794
dx.doi.org/10.1021/ja4048569 | J. Am. Chem. Soc. 2013, 135, 15794−15800

Journal of the American Chemical Society
Article
was synthesized by starting from commercially available toluene
and carbon tetrachloride, in reference to protocols reported in
the literature.17−23 The synthesis and isolation of the
unsymmetrical porphyrin in high yield often encounter
difficulties and require careful consideration in tuning the
molecular solubility and polarity. 4-Isopropoxybenzaldehyde
was selected to improve the solubilities and to discriminate the
polarities among the porphyryin mixtures so that efficient
isolation can be achieved. The condensation of pyrrole with 4-
isopropoxybenzaldehyde and 4-bromobenzaldehyde in a molar
ratio of 5:4:1 using the Alder−Longo method24 gave porphyrin
2 in 6% yield. 2 was then metallized by zinc acetate dihydrate to
yield zinc porphyrin 3 in a yield of 93%. Sonogashira coupling
of 3 with trimethylsilyacetylene (TMSA) afforded 4 in 70%
yield. Deprotection of the TMS group in 4 with tetrabuty-
lammonium fluoride (TBAF) produced 5 in 82% yield. The
cross-coupling reaction of 5 with Tpd-I furnished 6 in 63%
yield. Subsequent hydrolysis under acidic conditions to strip
both the acetyl group and Zn(II) at the same time produced
target free-base porphyrin molecule 1 with three mercapto-
methyl anchor groups in 84% yield. All new compounds (1−6)
were purified by column chromatography on silica gel and
characterized using various spectroscopic methods including 1H
and 13C NMR spectroscopy and ESI- or MALDI-TOF mass
spectrometry. More details on the synthesis can be found in the
Experimental Section. The thermal analysis data of 1 can be
found in the Supporting Information (SI).
Figure 1. Schematic configuration of 1 on Au(111) and localized
electrical excitation from a nanotip.
a
Scheme 1. Synthesis of Target Molecule 1
The luminescence property of the molecule thus designed
and synthesized is our primary concern and needs to be
checked before proceeding to the next step. Both steady-state
and transient spectral measurements were performed in order
to examine the influence of chemical modifications on the
optical properties of synthesized molecules. As shown in Figure
2a,b, both the UV−vis absorption and steady-state photo-
luminescence (PL) spectra of 1 in CH2Cl2 solution show
characteristic absorption and emission bands similar to those of
the tetraphenyl porphyrin moiety,4,10 also revealing an optical
band gap of ∼1.9 eV. In the characteristic double Q-band
a
(a) AlCl3, CCl4, 0 °C, 1 h, rt, 2 h; AcCl, 80 °C, 2 h, 25%. (b) aniline,
N2, 200 °C, 6 h, 87%. (c) NaNO2, HCl(aq) (6.1 M solution), EtOH, 0
°C, 1 h; KI, 0 °C, 1 h, rt, 19 h, 54%. (d) NBS, BPO, CCl4, Ar, 90 °C, 3
days; KSAc, THF, reflux, 24 h, 39%. (e) EtCOOH, (EtCO)2O, Ar, 170
°C, 3 h, 6%. (f) Zn(OAc)2·2H2O, CHCl3−MeOH (3/1), Ar, 80 °C, 5
h, 93%. (g) TMSA, PdCl2(PPh3)2, PPh3, CuI, THF, TEA, Ar, 80 °C,
48 h, 70%. (h) TBAF, THF, rt, 1 h, 82%. (i) Pd2(dba)3, PPh3, toluene,
TEA, Ar, 30 °C, 7 days, 63%. (j) conc. H2SO4, CH2Cl2, MeOH, 0 °C,
2 h, 60 °C, 12 h, 84%.
the axial nanocavity plasmon (NCP), which is expected to
provide maximal emission enhancement.8,10,11,14,15 The opti-
mized structure of 1 has a height of ∼3.1 nm and a width of
∼2.3 nm as well as a spacer length of ∼1.5 nm.
However, the multifunctionalized complexity and the
relatively large size of target molecule 1 make the sample
preparation a nontrivial task, which includes both the synthesis
of high-purity materials in high yield and the deposition of
molecules on the metal surface with well-defined structures for
STM studies. The synthesis procedure for 1 is shown in
Scheme 1. The iodo-substituted tripodal rigid anchor (Tpd-I)
Figure 2. (a) UV−vis absorption and (b) PL spectra of 1 in CH2Cl2
solution (∼1.0 × 10−5 mol L−1). The dashed curve in plot a shows
magnified data of the Q-band absorption. (c, d) Time-resolved PL
decay trace for 1 in CH2Cl2 (∼1.0 × 10−8 mol L−1) at 660 and 726
nm, respectively, with pulsed-laser excitation at 405 nm. The black
lines are the raw data, and the red lines are the fitting curves.
15795
dx.doi.org/10.1021/ja4048569 | J. Am. Chem. Soc. 2013, 135, 15794−15800

Journal of the American Chemical Society
Article
emission associated with the π*−π transitions of the conjugate
porphyrin plane, the band at ∼660 nm is attributed to the
Qx(0,0) transition whereas the band at ∼726 nm is attributed
to the Qx(0,1) transition.4,16 More importantly, the time-
resolved fluorescence decay data in Figure 2c,d reveal clear
single-exponential decay behavior for both emission bands, with
lifetimes of ∼7.86 ns at 660 nm and ∼7.72 ns at 726 nm. These
lifetimes are very close to the monomer lifetime of ∼8.38 ns for
the tetraphenyl porphyrin,9 which suggests that the chemical
modifications do not generate additional nonradiative channels
that would otherwise quench the molecular fluorescence. The
designed self-decoupled porphyrin molecule thus appears to be
a good candidate as an emitter for STML studies.
For large multifunctionalized molecules such as 1, vacuum
sublimation methods are no longer suitable because the
molecules will decompose during heating. Instead, we
employed wet-chemistry methods to deposit 1 on Au(111).
In an effort to reduce contamination, we first performed the
deposition experiment in the load-lock chamber under a
nitrogen-gas atmosphere via short-time drop-casting of 1
(∼10−6 mol/L in CH2Cl2) onto the atomically cleaned
Au(111) surface.
Note that, apart from target molecules 1, there also exist
some unknown impurity molecules (with thiol-related frag-
ments as one possibility) on the surface according to the
presence of faint dotlike features between the dragonfly-like
structures. To remove these impurities and also to change the
adsorption configuration of 1 on Au(111), the sample was
annealed at ∼310 °C for 1 h. The STM image in Figure 3b
shows that the annealing procedure indeed removes most of
the impurities and the adsorption configuration of 1 is also
changed. The surface appears to be cleaner overall, as illustrated
in the inset of Figure 3b (more in Figure S2c of the SI),
showing both discernible herringbone structures on the bare
Au(111) substrate and a stripelike molecular assembly.
However, the higher-resolution image of Figure 3b indicates
that the stripelike assembly is composed of a practically two-
layer structure with isolated bright units on the top and
continuous impurity stripes underneath. These individual bright
units on the top, with a feature size of about 3 to 4 nm, are
arranged into different packing configurations. Some appear to
be single poorly resolved bright protrusions (representative site
A), and others form five-dot patterns (representative site B).
Although the exact adsorption configurations are unclear, these
bright features are believed to associate with the self-decoupled
porphyrin molecules of 1, as confirmed by subsequent
annealing and STML studies. As illustrated in Figure 3c,
further annealing at a higher temperature of ∼400 °C
completely removes the impurity molecules from the sample
surface. Target molecule 1 is also disassembled, but its
porphyrin fragment survives on the surface. The characteristic
four-lobed pattern of a tetraphenyl porphyrin unit10,26 emerges,
but with an additional tail that is likely to associate with the (4-
methylphenyl)ethynyl group (schematic inset in Figure 3c). In
combination with the thermogravimetric analysis data (details
in Figure S1 and Scheme S1 in the SI), the tripod appears to
break up around this temperature and then desorbs from the
surface.
As shown in Figure 3a, there are isolated multidot features on
the STM image of the as-prepared sample, which, for the sake
Figure 3. (a) STM image of 1 as prepared via drop-casting on
Au(111) (2.8 V, 10 pA, 25 × 25 nm2). The inset shows the Au(111)
surface before deposition (2.5 V, 100 pA, 80 × 80 nm2). (b) STM
image upon annealing at ∼310 °C for ∼1 h (2.8 V, 20 pA, 30 × 30
nm2). The inset is an image of a larger area (2.8 V, 5 pA, 75 × 75 nm2)
with a discernible herringbone feature on the bare Au surface. (c)
High-resolution image upon further annealing at ∼400 °C for ∼1 h
(2.5 V, 10 pA, 12 × 12 nm2).
Achieving intrinsic molecular electroluminescence from a
single porphyrin molecule is the primary goal in this work.
Unlike the absence of molecular fluorescence for flat-lying 1 in
the as-prepared sample (Figure 3a), the STML spectrum shown
in Figure 4a above site A (Figure 3b) exhibits porphyrin-
specific emission of Qx(0,0) and Qx(0,1) at ∼1.88 and ∼1.74
of convenience, could each be visualized as a “dragonfly-like”
structure with a length of ∼3.3 nm and a width of ∼2.5 nm.
These dimensions are very close to those estimated for
designed self-decoupled porphyrin molecule 1. Each isolated
structure can be approximately considered to consist of four
bright dots. The front three-dot pattern may correspond to the
three 4-isopropoxyphenyl groups attached to the porphyrin
core, whereas the bright dot behind might be assigned to one
leg of the tripod with the other two legs bonded to the
substrate, as illustrated in the schematic molecular orientation
in Figure 3a (more in the SI). All of these topographical
features strongly suggest that each isolated dragonfly-like
structure corresponds to one target molecule 1. Nevertheless,
its adsorption configuration appears to lie flat on the Au(111)
surface rather than standing up with the desired tripod attached
to the substrate.25 The STML measurements on these lying-
down molecules exhibit no molecule-specific emission,
presumably owing to the quenching of fluorescence originating
from insufficient decoupling, misalignment of energy levels at
the interface, or the mismatch between the molecular transition
dipoles and the axial NCP or a mix of these.
Figure 4. (a) STML spectra on representative sites A−D in Figure 3b
(2.5 V, 50 pA, 2 min). (b) Bias dependence of STML spectra on site A
(30 pA, 1 min). The STML spectrum on Au(111) is also plotted as a
reference (1.8 V, 30 pA, 1 min). The intensities of all STML spectra
are normalized by both currents and the acquisition time. Spectra are
offset for clarity.
15796
dx.doi.org/10.1021/ja4048569 | J. Am. Chem. Soc. 2013, 135, 15794−15800

Journal of the American Chemical Society
Article
eV, respectively. By contrast, the STML spectrum above the
porphyrin molecules at site B does not exhibit any molecule-
specific fluorescence but only suppressed tip-induced plasmonic
emission from the metal substrate.27−30 The spectral feature is
very similar to those acquired on both the impurity layer (site
C) and the bare Au surface (site D). The suppression effect of
the emission intensity is qualitatively scaled to the topographic
height (i.e., the gap distance between the tip and metal
substrate).28,30 The sharp distinction of the STML spectra
between sites A and B for target molecule 1 suggests that the
molecule at site B still lies flat on the surface, whereas the
porphyrin core plane at site A is likely to orient to some extent
along the tip axial direction. The latter could provide a
considerable vertical component of the transition dipole to
couple effectively with the NCP to produce molecular light in
the far field. In other words, the molecules at site A are likely
either to stand up as designed or to at least be tilted on the
substrate, as also supported by the topographic information
with the apparent height of site A being higher than that of site
B (height profile details in the SI). It is worth pointing out that
the additional emission band around 1.59 eV on site A (Figure
4a) can be assigned to the Qx(0,2) transition. This low-energy
emission band is usually invisible in conventional PL measure-
ments and represents one unique characteristic of STM-
induced molecular electroluminescence defined by nanoscale
plasmonic environments. Its anomalous occurrence suggests
the presence of a strong emission enhancement by resonant
NCP fields10 because the tip-induced plasmonic modes on the
bare Au (site D in Figure 4a) are very broad, with considerable
intensities extending into the low-energy regime. It should be
noted that the percentage of fluorescent spots in Figure 3b for
the annealed drop-casted sample is very small, less than 5%, and
most of them are located at the ends of the bright stripes. For
the same tip status, similar spectral shapes could be obtained
for similar emitting spots although the intensity might vary.
Further evidence for molecular fluorescence in site A is the
nearly constant peak position for each emission band as a
function of the excitation voltage (Figure 4b) because the
radiative decay of the excited molecule is associated with a
given HOMO (the highest-occupied molecular orbital)−
LUMO (the lowest-unoccupied molecular orbital) gap. An
onset voltage at ∼1.9 V is observed, which is about the same as
the optical band gap of the molecule. Above the threshold
voltage, the STML intensity rises with increasing bias voltage
(up to about 2.8 V) owing to both the growing number of
energetically allowed channels that could contribute to photon
emission and the enhanced electromagnetic field underneath
the tip. Note that the broad emission band at ∼1.8 V in Figure
4b for the site-A molecules originates from the NCP emission
alone because its spectral profile is very similar to that acquired
on the bare Au surface, though dramatically suppressed as a
result of the greatly increased gap distance.
Au(111) in the liquid, we have examined another wet-chemistry
method by immersing the Au(111) substrate in the 10−6 mol/L
CH2Cl2 solution of 1 for 10 h at room temperature. (See the SI
for more details on the influence of concentration and
immersion time on the orientational control of the molecules
on Au.)
In contrast to the STM image for the drop-casted sample
(Figure 3a), we no longer observe dragonfly-like structures
associated with the flat-lying molecules but instead many
scattered bright protrusions with a dot dimension in the range
of about 3−5 nm (Figure 5a). Although the surface still
Figure 5. (a) STM image of 1 by immersing Au(111) in the CH2Cl2
solution of 1 for 10 h (2.8 V, 10 pA, 60 × 60 nm2). (b) Bias
dependence and polarity dependence of STML spectra on top of the
molecules(s) inside the blue circle (50 pA, 1 min). The intensities of
all STML spectra are normalized by both the currents and acquisition
time. The spectra are offset for clarity.
contains unknown impurity molecules, the step-edge feature of
the Au(111) surface is clearly discernible, together with a small
portion of bare, exposed Au. All of these observations suggest
that the bright protrusion may correspond to either a single
molecule of 1 or its aggregate (two to three molecules)
standing up directly on the substrate. Such a hypothesis is
confirmed by subsequent STML experiments, which indicates
that about 60−70% of the bright spots can produce strong
molecular electroluminescence. Figure 5b shows the represen-
tative STML spectra acquired (e.g., on the dot inside the circle
in Figure 5a). Not only is the porphyrin-specific double Q-band
emission clearly observed, but also the normalized emission
intensity of the strongest peak is larger than that obtained on
site A of the annealed drop-casted sample (Figures 3b and 4b).
The stronger emission capability for the long-time immersion
sample is also supported by the quantum efficiency measure-
ments. A raw photon intensity of ∼460 counts per second (cps)
was detected by the single photon avalanched photodiode
(SPAD) at 10 pA on the emitting molecules in Figure 5a,
which, upon calibration of the photon collection and detection
efficiency, gives rise to a quantum efficiency of ∼2 × 10−4
photons/electron. Whereas on the emitting molecules at site A
for the annealed drop-casted sample in Figure 3b, the quantum
efficiency is estimated to be ∼4 × 10−5, about 1 order of
magnitude weaker. All of these observations of topographic and
spectral features as well as emission intensities suggest that
molecule 1 is likely to stand up in the long-time immersion
sample, allowing the molecular transition dipole to couple
efficiently with the NCP modes to generate strong emission.
Note that the low-energy Qx(0,2) emission at around 1.59 eV
in Figure 4 does not show up here because the NCP resonance
mode for the tip used for the STML measurements in Figure 5
peaks at around 660 nm, having little plasmonic intensity above
Although molecule-specific emission can be realized on the
above drop-casted sample upon annealing, the probability of
obtaining molecular electroluminescence is very low and
difficult to control. The unfavorable flat-lying adsorption
configuration of 1 in the drop-casted sample (Figure 3a) may
pertain to the deposition procedure. When the solution is drop-
casted on the Au(111) surface under nitrogen flow in the load-
lock chamber, the solvent vaporizes too fast to provide
sufficient time for all three arms of the tripod to anchor on
Au(111) in the liquid state. Therefore, to ensure that molecule
1 has sufficient time to align itself to make better contact with
15797
dx.doi.org/10.1021/ja4048569 | J. Am. Chem. Soc. 2013, 135, 15794−15800

Journal of the American Chemical Society
Article
750 nm. As a result, the conditions for the emission
enhancement of the Qx(0,2) band cannot be created. These
observations demonstrate again the spectral shaping capability
of the resonant NCP modes on molecular fluorescence thanks
to greatly enhanced radiative decay rates.10
spectra simultaneously using the same tip. Some typical dI/dV
data associated with the standing-up molecules are given in the
SI, which shows the occurrence of well-defined LUMO states
that are peaked at around +2.2(2) V. More importantly, the
onset position of the LUMO states in the differential
conductance curve appears to correlate approximately with
the onset voltage of molecular electroluminescence (more
details in the SI).
Figure 5b also reproduces the onset voltage of molecular
electroluminescence at ∼1.9 V and increased emission intensity
along with the increase in bias voltage. Another important
spectral feature in Figure 5b is the highly asymmetric
dependency on the bias polarity (e.g., at 2.5 V). Unlike the
ambipolar operation of molecular electroluminescence for
multimonolayer molecules,4,10 the molecule in the present
system fluoresces only at positive bias voltages. The origin of
such unipolar performance is still not completely clear but is
believed to associate with both the energy-level alignment at
the molecular interface and the junction asymmetry defined by
the tip−molecule−substrate, which could result in different
distributions of voltage drops across the STM junction.31−33
The junction between the tip and molecule is a vacuum barrier,
and the contact between the molecule and substrate through
the S−Au bonds is quite strong (Figure 6a). Although the
Furthermore, the unipolar performance also suggests that the
NCP field generated by tunneling electrons seems to play a
minor role in the excitation of molecules here. Otherwise,
because plasmonic excitation is bias-polarity irrelevant,
molecular fluorescence should be observed for both polarities.
The exact reason for the inefficient excitation of molecules by
plasmons in the present system is still unclear but may pertain
to the very low efficiency of electron energy being inelastically
converted to the NCP emission in the tunneling regime (less
than 10−3 photons per electron).27,34−36 As a result, the
excitation probability of molecules by plasmons is much smaller
compared to the direct excitation via elastic electron injection.
However, it should be pointed out that resonant NCP modes
are believed to play an important role in the emission
enhancement through improving the radiative decay
rates.8,10,14,15 Figure 6 illustrates such an excitation and
emission process schematically: when both the HOMO and
LUMO levels fall inside the electrochemical potential window
defined by the bias voltage, the molecule is excited
predominantly by hot electron injection;4,5 the excited
molecule then decays back to the ground state via Franck−
Condon π*−π transitions with the radiative rate enhanced by
the nanocavity plasmons resonant with the selected channels,10
producing enhanced molecular electroluminescence.
CONCLUSIONS
■
We have successfully designed and synthesized a tripodal
porphyrin molecule that could produce molecular electro-
luminescence on the single-molecule scale. Dipping the Au
substrate into the molecular solution for a sufficiently long time
is found to be important for the molecule to stand up as
desired, with the tripod anchored to the substrate and the
porphyrin plane aligned along the tip axial direction. Moreover,
the designed self-decoupled molecule makes the tip−
molecule−substrate junction highly asymmetric, which enables
the observation of a new molecular electroluminescence
phenomenon such as the unipolar performance. Our results
help us to gain insight into the dominant excitation mechanism
via hot-electron injection and emission enhancement through
efficient plasmon-exciton coupling. These findings, together
with the self-decoupling design to suppress fluorescence
quenching, may open up a new way to realize electrically
driven single-molecule light sources and to improve the
molecule−electrode interface in organic light-emitting diodes.
Figure 6. (a) Schematic diagrams showing molecular excitation via
hot-electron injection for the positive-sample-bias situation, followed
by plasmon-enhanced emission. (b) Molecular fluorescence via a
Franck−Condon transition from the excited state (S1) to the ground
state (S0).
build-in spacer may decouple the interaction of the HOMO
and LUMO states of the porphyrin moiety with the substrate to
some extent, these frontier molecular orbitals may still be more
or less pinned to the substrate and the voltage drop at the
molecule−substrate interface could be relatively small. Because
the optical band gap of molecule 1 is about 1.9 eV, the
occurrence of molecular electroluminescence at positive bias
above ∼1.9 V suggests that the HOMO level of the molecule
under such bias has to align with the Fermi level of the Au
substrate so that the HOMO states can be partially emptied to
generate holes for radiative combinations with electrons from
the LUMO states. However, the absence of molecular
electroluminescence at negative bias suggests that in this case
it is very difficult, if not impossible, to lift the Fermi surface of
the substrate to reach the LUMO level for electron injection
within a reasonable bias range that could survive the molecule.
In principle, information about such energy-level alignment
could be obtained from the differential conductance (dI/dV)
measurements. Nevertheless, because of the relatively “dirty”
nature of the wet-chemistry sample preparation method for
such a complex multifunctionalized molecule and its relatively
large height, it is quite demanding experimentally to obtain
well-defined data for topography, dI/dV−V curves, and STML
EXPERIMENTAL SECTION
■
Synthesis of 5-(4-Bromophenyl)-10,15,20-tris-(4-
isopropoxyphenyl)porphyrin 2. To a solution containing 4-
isopropoxybenzaldehyde (18.71 g, 114 mmol) and 4-bromobenzalde-
hyde (5.29 g, 28.6 mmol) in 140 mL of propionic acid was added 60
mL of propionic anhydride. After stirring under argon at room
temperature for 30 min, the solution was heated to 170 °C and
allowed to reflux under the protection of argon with vigorous stirring,
and then freshly distilled pyrrole (10 mL, 141 mmol) in propionic acid
(10 mL) was added dropwise. After the addition, the reaction mixture
was stirred for another 3 h at 170 °C, and then anhydrous ethanol (40
15798
dx.doi.org/10.1021/ja4048569 | J. Am. Chem. Soc. 2013, 135, 15794−15800

Journal of the American Chemical Society
Article
mL) was added under vigorous stirring. The mixture was cooled to
room temperature and filtered, and the solid product was washed
repeatedly with ethanol until the rinsing solution was no longer dark.
The filter cake was dissolved in CH2Cl2 and chromatographed on silica
gel using CH2Cl2−petroleum ether (1/1) as the eluent to afford
phenyl]methane (268.7 mg, 0.38 mmol) in dry toluene−triethylamine
(18 mL, 5/1) was purged with argon for 30 min. The coupling was
initiated by adding PPh3 (113.0 mg, 0.43 mmol) followed by
Pd2(dba)3 (114.5 mg, 0.13 mmol), and the reaction mixture was
then stirred at 30 °C for 7 days until thin-layer chromatography
(TLC) monitoring indicated the formation of new spots and the
disappearance of 5. The crude reaction mixture was purified by silica
gel chromatography using petroleum ether/dichloromethane (1/2) as
1
purple crystalline porphyrin 2 (1.51 g, 6%). H NMR (400 MHz,
CDCl3) δ 8.89 (d, J = 4.8 Hz, 2H), 8.88 (s, 4H), 8.78 (d, J = 4.8 Hz,
2H), 8.09−8.04 (m, 8H), 7.84 (d, J = 8.2 Hz, 2H), 7.23−7.21 (m,
6H), 4.81 (heptet, J = 6.0 Hz, 1H), 4.80 (heptet, J = 6.0 Hz, 2H), 1.53
(d, J = 6.0 Hz, 18H), −2.75 (bs, 2H, NH). 13C NMR (100 MHz,
CDCl3) δ 157.9, 141.5, 136.0, 135.8, 134.5, 134.4, 130.0, 122.5, 120.5,
120.3, 118.1, 114.1, 70.3, 22.5. MALDI-TOF-MS m/z: [M+] calcd for
C53H4779BrN4O3 866.2832, found 866.2059.
1
the eluent to furnish porphyrin 6 (230.7 mg, 63% yield). H NMR
(400 MHz, CDCl3) δ 8.99 (d, J = 4.6 Hz, 2H), 8.98 (s, 4H), 8.92 (d, J
= 4.6 Hz, 2H), 8.18 (d, J = 8.1 Hz, 2H), 8.08 (d, J = 8.4 Hz, 6H), 7.88
(d, J = 8.1 Hz, 2H), 7.54 (d, J = 8.5 Hz, 2H), 7.24−7.21 (m, 6H),
7.15−7.03 (m, 14H), 4.88−4.79 (m, 3H), 3.96 (s, 6H), 2.21 (s, 9H),
1.55 (d, J = 6.0 Hz, 18H). 13C NMR (100 MHz, CDCl3) δ 195.3 (C
O), 157.7, 150.8, 150.7, 150.6, 149.9, 147.1, 145.4, 143.3, 143.2, 135.6,
135.4, 135.13, 135.10, 134.6, 132.3, 132.2, 132.1, 131.6, 131.3, 131.2,
131.1, 130.0, 129.0, 128.9, 128.4, 128.2, 127.4, 122.5, 121.3, 121.1,
119.9, 114.0, 90.3, 89.8, 70.3, 64.4, 33.0, 30.3, 22.5. MALDI-TOF-MS
m/z: [M+] calcd for C89H76N4O6S3Zn 1456.4218, found 1456.4545.
Synthesis of Porphyrin 1. Porphyrin 6 (43.2 mg, 0.03 mmol) was
placed in a 250 mL three-necked round-bottomed flask and dissolved
with degassed CH2Cl2 (56 mL) and degassed MeOH (28 mL).
Concentrated H2SO4 (5.6 mL) was added dropwise to the mixture at 0
°C. The mixture was then gradually warmed to 60 °C with stirring.
After stirring for 12 h, the reaction was quenched by the addition of
water, and the organic layer was separated. The aqueous layer was
extracted with CH2Cl2, and the combined organic layer was washed
with brine and dried over Na2SO4. After the removal of the solvent
under reduced pressure, the residue was purified by column
chromatography on silica gel with CH2Cl2 as the eluent to give target
molecule 1 (31.6 mg, 84% yield). 1H NMR (400 MHz, CDCl3) δ 8.90
(d, J = 4.8 Hz, 2H), 8.89 (s, 4H), 8.83 (d, J = 4.8 Hz, 2H), 8.19 (d, J =
8.1 Hz, 2H), 8.09 (d, J = 8.4, 6H), 7.89 (d, J = 8.0 Hz, 2H), 7.56 (d, J
= 8.4 Hz, 2H), 7.29−7.19 (m, 20H), 4.85 (heptet, J = 6.1 Hz, 3H),
3.74 (d, J = 7.5 Hz, 6H), 1.79 (t, J = 7.5 Hz, 3H), 1.56 (d, J = 6.1 Hz,
18H), −2.74 (bs, 2H, NH). 13C NMR (100 MHz, CDCl3) δ 157.9,
147.3, 145.33, 145.27, 142.6, 139.0, 135.9, 134.8, 134.5, 134.4, 131.4,
131.23, 131.18, 130.1, 127.5, 122.8, 121.1, 120.5, 120.2, 118.9, 114.1,
90.5, 89.8, 70.3, 64.5, 28.6, 22.5. ESI-MS m/z: [M + H+] calcd for
C83H72N4O3S3 1268.4767, found 1269.4840.
Synthesis of Zinc(II)-5-(4-bromophenyl)-10,15,20-tris-(4-
isopropoxyphenyl)porphyrin 3. A mixture of 5-(4-bromophen-
yl)-10,15,20-tris-(4-isopropoxyphenyl)porphyrin 2 (1.03 g, 1.19
mmol) and Zn(OAc)2·2H2O (2.80 g, 12.7 mmol) was dissolved in
chloroform−CH3OH (60 mL, 3/1), and then the mixture was allowed
to reflux under argon for 5 h. After the removal of the solvents, the
residue was chromatographed on a silica gel column (CH2Cl2−
1
petroleum ether, 1/1) to give porphyrin 3 (1.02 g, 93% yield). H
NMR (400 MHz, CDCl3) δ 8.99−8.96 (m, 6H), 8.86 (d, J = 4.7 Hz,
2H), 8.07−8.02 (m, 8H), 7.81 (d, J = 8.2 Hz, 2H), 7.20−7.16 (m,
6H), 4.77 (heptet, J = 6.0 Hz, 1H), 4.76 (heptet, J = 6.0 Hz, 2H), 1.51
(d, J = 6.0 Hz, 18H). 13C NMR (100 MHz, CDCl3) δ 157.7, 150.9,
150.8, 150.7, 149.9, 142.1, 135.9, 135.6, 135.1, 135.0, 132.4, 132.3,
132.2, 131.5, 129.8, 122.2, 121.5, 121.3, 119.2, 114.0, 70.3, 22.5.
MALDI-TOF-MS m/z: [M+] calcd for C53H4579BrN4O3Zn 928.1967,
found 928.0060.
Synthesis of TMS-Protected Zinc(II)-5-(4-ethynylphenyl)-
10,15,20-tris-(4-isopropoxyphenyl)porphyrin 4. Air was re-
moved from a single-necked flask containing a mixture of 3 (433.2
mg, 0.47 mmol), Pd(PPh3)2Cl2 (84.1 mg, 0.12 mmol), CuI (46.5 mg,
0.24 mmol), and PPh3 (60.0 mg, 0.23 mmol) in THF−Et3N (35 mL,
6/1) by argon bubbling for 30 min. Then trimethylsilylacetylene (450
μL, 3.19 mmol) was added, and the mixture was allowed to stir under
argon at room temperature for 10 min. Thereafter, the flask was sealed
and heated at 80 °C for 48 h. After cooling to room temperature, the
solvent was evaporated, and the crude product was purified on a silica
gel column (CH2Cl2−petroleum ether, 1/1) to provide TMS-
1
protected porphyrin 4 (308.5 mg, 70% yield). H NMR (400 MHz,
Theoretical Calculation. The geometry of 1 was calculated at the
B3LYP/6-31G(d) level with Gaussian 09W without a consideration of
the effect of the substrate.
CDCl3) δ 8.97 (d, J = 4.7 Hz, 2H), 8.96 (s, 4H), 8.86 (d, J = 4.7 Hz,
2H), 8.12 (d, J = 8.2 Hz, 2H), 8.054 (d, J = 8.6 Hz, 2H), 8.047 (d, J =
8.6 Hz, 4H), 7.85 (d, J = 8.2 Hz, 2H), 7.23 (d, J = 8.6 Hz, 6H), 4.84
(heptet, J = 6.0 Hz, 3H), 1.56 (d, J = 6.0 Hz, 18H), 0.38 (s, 9H). 13C
NMR (100 MHz, CDCl3) δ 157.8, 150.8, 150.73, 150.68, 149.9, 143.5,
135.6, 135.11, 135.07, 134.5, 132.4, 132.2, 132.1, 131.6, 130.4, 122.4,
121.4, 121.2, 119.9, 114.0, 105.4, 95.4, 70.3, 22.5, 0.3. MALDI-TOF-
MS m/z: [M+] calcd for C58H54N4O3SiZn 946.3257, found 946.2683.
Synthesis of Zinc(II)-5-(4-ethynylphenyl)-10,15,20-tris-(4-
isopropoxyphenyl)porphyrin 5. Porphyrin 4 (308.5 mg, 0.33
mmol) was dissolved in distilled THF (30 mL) and stirred under an
argon atmosphere for 15 min. Then tetrabutylammonium fluoride
(1.69 g, 6.48 mmol) was added. The mixture was stirred under argon
for 1 h and then diluted with dichloromethane (100 mL) and washed
twice with water. The organic phase was collected and dried over
anhydrous sodium sulfate. The solvent was removed under reduced
pressure, and the crude product was purified on a silica gel column
(CH2Cl2−petroleum ether, 1/1) to yield porphyrin 5 (235.0 mg, 82%
Acquisition of Photoluminescence Spectra and Measure-
ment of Photoluminescence Lifetimes. The acquisition of PL
spectra and the measurement of fluorescence lifetimes were carried out
on a home-built optical setup. Steady-state spectra were recorded with
a liquid-nitrogen-cooled charge-coupled-device (CCD) spectrometer
(Princeton Instruments) under the excitation of light at 405 nm. The
time-correlated single-photon counting technique (Edinburgh Instru-
ments) was used to determine the fluorescence lifetimes. The time
resolution was ∼8 ps with the use of a microchannel plate
photomultiplier tube and a pulse picosecond diode laser at 405 nm
(Hamamatsu). A cutoff filter at 470 nm was used to block the
excitation light. The time-resolved PL curves were fitted with an
exponential function.
Molecular Deposition on Au(111). After being synthesized and
purified, molecule 1 was immediately dissolved in CH2Cl2 at a
concentration of ∼1.0 × 10−6 M.
1
yield). H NMR (400 MHz, CDCl3) δ 9.00 (d, J = 4.7 Hz, 2H), 8.99
Wet-Chemistry Method via Short-Time Drop-Casting. The drop-
casted sample was prepared by dripping 10 μL of a solution of 1 onto
the atomically cleaned Au(111) surface in the load-lock chamber
under a nitrogen-gas atmosphere. The solvent evaporated within
seconds, and the chamber was immediately pumped under vacuum.
The sample was then transferred to the ultrahigh-vacuum (UHV)
observation chamber for STM investigations at low temperature (∼80
K).
(s, 4H), 8.90 (d, J = 4.7 Hz, 2H), 8.18 (d, J = 8.2 Hz, 2H), 8.10 (d, J =
8.5 Hz, 2H), 8.09 (d, J = 8.5 Hz, 4H), 7.88 (d, J = 8.2 Hz, 2H), 7.25
(d, J = 8.4 Hz, 6H), 4.86 (heptet, J = 6.0 Hz, 3H), 3.30 (s, 1H), 1.57
(d, J = 6.0 Hz, 18H). 13C NMR (100 MHz, CDCl3) δ 157.8, 150.9,
150.74, 150.68, 149.9, 143.8, 135.6, 135.1, 135.0, 134.5, 132.4, 132.24,
132.16, 131.6, 130.5, 121.5, 121.4, 121.3, 119.8, 114.0, 84.0, 78.2, 70.3,
22.5. MALDI-TOF-MS m/z: [M+] calcd for C55H46N4O3Zn 874.2861,
found 874.2104.
Wet-Chemistry Method via Long-Time Immersion. The freshly
prepared Au(111) substrate was immersed in the solution for 10 h at
room temperature. Then the substrate was removed from the solution,
Synthesis of Porphyrin 6. A solution of porphyrin 5 (219.8 mg,
0.25 mmol) and 1-(4-iodophenyl)-1,1,1-tris[4-(S-acetylthiomethyl)-
15799
dx.doi.org/10.1021/ja4048569 | J. Am. Chem. Soc. 2013, 135, 15794−15800

Journal of the American Chemical Society
Article
rinsed thoroughly with CH2Cl2 10 times, and dried under a stream of
nitrogen gas before being loaded into the UHV chamber for STM
investigations at low temperature (∼80 K). The influence of
concentration and immersion time on the adsorption configuration
of the target molecule on Au is briefly described in the SI.
(9) Zhang, X.-L.; Chen, L.-G.; Lv, P.; Gao, H.-Y.; Wei, S.-J.; Dong,
Z.-C.; Hou, J. G. Appl. Phys. Lett. 2008, 92, 223118.
(10) Dong, Z. C.; Zhang, X. L.; Gao, H. Y.; Luo, Y.; Zhang, C.; Chen,
L. G.; Zhang, R.; Tao, X.; Zhang, Y.; Yang, J. L.; Hou, J. G. Nat.
Photonics 2010, 4, 50.
Scanning Tunneling Microscope-Induced Luminescence.
The experiments were performed with a low-temperature UHV-
STM (Unisoku) at a base pressure of ∼8 × 10−11 mbar at ∼80 K,
operated in the constant-current−topographic mode with the sample
biased. The Au(111) substrate was prepared by the thermal
evaporation of gold (∼150 nm thick) on freshly cleaved mica and
cleaned in UHV by cycles of argon ion sputtering. Electrochemically
etched silver (Ag) tips were used to perform STM imaging and STML
measurements, taking advantage of silver in producing strong
plasmonic fields.35
Photons emitted from the tunnel junction were collected with a lens
inside the UHV chamber and then refocused into an optical fiber by
another lens placed outside the UHV chamber. The collected light was
guided into a grating spectrometer and analyzed by a cooled CCD
detector (Princeton Instruments) for spectral information.10,35,36
Photon intensities, measured by the SPAD (PerkinElmer, ∼72%
detection efficiency of around 660 nm), were corrected for the
detection efficiency of the SPAD and the collection efficiency of both
the lens system (∼10%) and the beamsplitter (50%).36 Nevertheless,
luminescence spectra presented in the article were not corrected for
the wavelength-dependent sensitivity of photon collection and
detection systems.
(11) Kabakchiev, A.; Kuhnke, K.; Lutz, T.; Kern, K. Chem. Phys.
Chem. 2010, 11, 3412.
(12) Matino, F.; Schull, G.; Kohler, F.; Gabutti, S.; Mayor, M.;
̈
Berndt, R. Proc. Natl. Acad. Sci. U.S.A. 2011, 108, 961.
(13) Schneider, N. L.; Matino, F.; Schull, G.; Gabutti, S.; Mayor, M.;
Berndt, R. Phys. Rev. B 2011, 84, 153403.
(14) Mills, D. L. Phys. Rev. B 2002, 65, 125419.
(15) Kuhn, S.; Mori, G.; Agio, M.; Sandoghdar, V. Mol. Phys. 2008,
̈
106, 893.
(16) Milgrom, L. R. The Colour of Life: An Introduction to the
Chemistry of Porphyrin and Related Compounds; Oxford University
Press: New York, 1997.
(17) Davis, G. L.; Hey, D. Y.; Williams, G. H. J. Chem. Soc. 1956,
4397.
(18) Heim, C.; Affeld, A.; Nieger, M.; Vogtle, F. Helv. Chim. Acta
̈
1999, 82, 746.
(19) Zhu, L.; Tang, H.; Harima, Y.; Yamashita, K.; Hirayama, D.;
Aso, Y.; Otsubo, T. Chem. Commun. 2001, 1830.
(20) Hirayama, D.; Takimiya, K.; Aso, Y.; Otsubo, T.; Hasobe, T.;
Yamada, H.; Imahori, H.; Fukuzumi, S.; Sakata, Y. J. Am. Chem. Soc.
2002, 124, 532.
(21) Wei, L.; Padmaja, K.; Youngblood, W. J.; Lysenko, A. B.;
Lindsey, J. S.; Bocian, D. F. J. Org. Chem. 2004, 69, 1461.
(22) Sakata, T.; Maruyama, S.; Ueda, A.; Otsuka, H.; Miyahara, Y.
Langmuir 2007, 23, 2269.
(23) Ie, Y.; Hirose, T.; Yao, A.; Yamada, T.; Takagi, N.; Kawai, M.;
Aso, Y. Phys. Chem. Chem. Phys. 2009, 11, 4949.
(24) Adler, A. D.; Longo, F. R.; Finarelli, J. D.; Goldmacher, J.;
Assour, J.; Korsakoff, L. J. Org. Chem. 1967, 32, 476.
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental procedures for the synthesis of Tpd-I, thermog-
ravimetric analysis, additional STM and tunneling spectral
information, Cartesian coordinates of molecule 1, and NMR
spectra of synthesized compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
(25) Schramm, A.; Stroh, C.; Dossel, K.; Lukas, M.; Fischer, M.;
̈
Schramm, F.; Fuhr, O.; v. Lohneysen, H.; Mayor, M. Eur. J. Inorg.
Chem. 2013, 2013, 70.
̈
AUTHOR INFORMATION
Corresponding Authors
gwang@ustc.edu.cn
(26) Jung, T. A.; Schlittler, R. R.; Gimzewski, J. K. Nature 1997, 386,
696.
(27) Berndt, R.; Gimzewski, J. K.; Johansson, P. Phys. Rev. Lett. 1991,
67, 3796.
(28) Tao, X.; Dong, Z. C.; Yang, J. L.; Luo, Y.; Hou, J. G.; Aizpurua,
J. J. Chem. Phys. 2009, 130, 084706.
(29) Zhang, Y.; Geng, F.; Gao, H. Y.; Liao, Y.; Dong, Z. C.; Hou, J.
G. Appl. Phys. Lett. 2010, 97, 243101.
(30) Geng, F.; Zhang, Y.; Yu, Y.; Kuang, Y.; Liao, Y.; Dong, Z.; Hou,
J. Opt. Express 2012, 20, 26725.
■
zcdong@ustc.edu.cn
Author Contributions
S.-E.Z. and Y.-M.K. contributed equally.
Notes
The authors declare no competing financial interest.
(31) Zahid, F.; Paulsson, M.; Datta, S. Electrical Conduction through
Molecules. In Advanced Semiconductors and Organic Nano-Techniques;
Morkoc, H., Ed.; Academic Press: New York, 2003; Vol. III, pp 1−41.
(32) Wu, S. W.; Nazin, G. V.; Chen, X.; Qiu, H. X.; Ho, W. Phys. Rev.
Lett. 2004, 93, 236802.
ACKNOWLEDGMENTS
■
We thank Andre Gourdon for helpful discussions on the
molecular design. This work is supported by NBRPC
(2011CB921402), CAS (XDB01020000), and NSFC
(91021004 and 10974186).
(33) Niquet, Y. M.; Delerue, C.; Allan, G. Phys. Rev. B 2002, 65,
165334.
(34) Johansson, P. Phys. Rev. B 1998, 58, 10823.
(35) Zhang, C.; Gao, B.; Chen, L. G.; Meng, Q. S.; Yang, H.; Zhang,
R.; Tao, X.; Gao, H. Y.; Liao, Y.; Dong, Z. C. Rev. Sci. Instrum. 2011,
82, 083101.
(36) Chen, L. G.; Zhang, C.; Zhang, R.; Zhang, X. L.; Dong, Z. C.
Rev. Sci. Instrum. 2013, 84, 066106.
REFERENCES
■
(1) Coombs, J. H.; Gimzewski, J. K.; Reihl, B.; Sass, J. K.; Schlittler,
R. R. J. Microsc. 1988, 152, 325.
(2) Berndt, R.; Gaisch, R.; Gimzewski, J. K.; Reihl, B.; Schlittler, R.
R.; Schneider, W. D.; Tschudy, M. Science 1993, 262, 1425.
(3) Qiu, X. H.; Nazin, G. V.; Ho, W. Science 2003, 299, 542.
(4) Dong, Z. C.; Guo, X. L.; Trifonov, A. S.; Dorozhkin, P. S.; Miki,
K.; Kimura, K.; Yokoyama, S.; Mashiko, S. Phys. Rev. Lett. 2004, 92,
086801.
́
(5) Cavar, E.; Blum, M.-C.; Pivetta, M.; Patthey, F.; Chergui, M.;
̈
Schneider, W.-D. Phys. Rev. Lett. 2005, 95, 196102.
(6) Chance, R. R.; Prock, A.; Silbey, R. Adv. Chem. Phys. 1978, 37, 1.
(7) Barnes, W. L. J. Mod. Opt. 1998, 45, 661.
(8) Anger, P.; Bharadwaj, P.; Novotny, L. Phys. Rev. Lett. 2006, 96,
113002.
15800
dx.doi.org/10.1021/ja4048569 | J. Am. Chem. Soc. 2013, 135, 15794−15800