Stable hydrocarbon radicals are utilized as spin standards and prototype metal-free molecular magnets able to withstand ambient conditions. Our study presents experimental results obtained with submolecular resolution by scanning tunneling microscopy and spectroscopy from monomers and dimers of stable hydrocarbon π radicals adsorbed on the Au(111) surface at 7-50 K. We provide conclusive evidence of the preservation of the radical spin-1/2 state, aiming to establish α,γ-bisdiphenylene-β-phenylallyl (BDPA) on Au(111) as a novel Kondo system, where the impurity spin is localized in a metal-free π molecular orbital of a neutral radical state in gas phase preserved on a metal support.
Stable hydrocarbon radicals are utilized as spin standards and prototype metal-free molecular magnets able to withstand ambient conditions. Our study presents experimental results obtained with submolecular resolution by scanning tunneling microscopy and spectroscopy from monomers and dimers of stable hydrocarbon π radicals adsorbed on the Au(111) surface at 7-50 K. We provide conclusive evidence of the preservation of the radical spin-1/2 state, aiming to establish α,γ-bisdiphenylene-β-phenylallyl (BDPA) on Au(111) as a novel Kondo system, where the impurity spin is localized in a metal-free π molecular orbital of a neutral radical state in gas phase preserved on a metal support.
Since the first observations
by Gomberg,[1,2] stable
free radicals are desirable spin standards, polarizing agents, and
building blocks of carbon-based magnetic systems.[3,4] Purely
hydrocarbon stable radicals are gaining ground as versatile model
systems for the development of novel π-magnetic systems[3,5,6] along with other benzenoid compounds
like graphene flakes,[7] carbon nanotubes,[8] fullerenes,[9] and π-conjugated
polymers.[10] Recently, the real-space study
of single unpaired electron spins on the atomic scale has become a
reality with the help of electron-transport experiments in surface-supported
d metal impurities[11,12] and metalorganic molecules[13−15] inside the tunnel junction of a scanning tunneling microscope (STM).
Such experiments utilize the Kondo effect,[16,17] that is, the screening of a localized electron spin by delocalized
conduction electrons at sufficiently low temperatures, as experimental
proof of the spin system maintaining its unpaired electron.Here we report on a low-temperature STM study of a new type of
Kondo system based on metal-free stable hydrocarbon π radicals,
BDPA (α,γ-bisdiphenylene-β-phenylallyl)[18] (Figure 1a), that organize
by self-assembly into different 1-D nanostructures on a Au(111) surface
while maintaining their spin-1/2 state. Our results based on scanning
tunneling spectroscopy (STS) reveal the characteristic properties
of the surface Kondo effect for BDPA/Au(111) with submolecular resolution
for isolated BDPA monomers and dimers. Compared with the monomer with
a Kondo temperature, TK, close to 50 K,
an increase in TK is observed for dimers,
suggesting a stabilization of the Kondo state by the radical–radical
interaction.
Figure 1
(a) DFT-calculated isodensity representation
of the singly occupied
molecular orbital (SOMO) of the stable π radical BDPA (C33H21) from ref (19). (b) STM topographic image of 0.2 monolayers
of BDPA grown at 300 K on the Au(111) surface (+0.5 V, 50 pA, 7 K);
scale bar = 1 nm. (c,d) Spectroscopic dI/dV maps at different electron energies (6 × 6 nm2); overlaid contour lines (dashed) mark areas covered by BDPA
molecules as extracted from STM topographs; arrows indicate a surface
defect.
Experimental Methods
BDPA recrystallized
in benzene was thermally evaporated in ultrahigh
vacuum (UHV) from a quartz crucible at 383 K after thorough degassing
at 373 K. The single-crystal Au(111) surface was prepared by repeated
cycles of 0.5 keV Ar+ bombardment and annealing at 720
K. STM and STS experiments were carried out at 5–7 K employing
electrochemically etched W tips deoxidized by annealing in UHV. The
dI/dV signal was obtained from the
first-harmonic current signal detected by lock-in technique (0.5 to
2 kHz; 5–20 mV sinusoidal peak-to-peak voltage; average of
10 single spectra). Impurity and tip effects were minimized by careful
sample preparation and multiple tip formings between the dI/dV experiments, resulting in Au-coated
STM tips. Reliable tip performance was established by accurately
reproducing the dI/dV signature
of the Au(111) surface state from literature.[20] dI/dV spectra were recorded under
constant-height conditions, the spectroscopic images (dI/dV maps) were simultaneously recorded with constant-current
STM topographic imaging.(a) DFT-calculated isodensity representation
of the singly occupied
molecular orbital (SOMO) of the stable π radical BDPA (C33H21) from ref (19). (b) STM topographic image of 0.2 monolayers
of BDPA grown at 300 K on the Au(111) surface (+0.5 V, 50 pA, 7 K);
scale bar = 1 nm. (c,d) Spectroscopic dI/dV maps at different electron energies (6 × 6 nm2); overlaid contour lines (dashed) mark areas covered by BDPA
molecules as extracted from STM topographs; arrows indicate a surface
defect.
Results and Discussion
BDPA Self-Assembly on Au(111)
Deposition of submonolayer
coverages of BDPA onto Au(111) at 300 K leads to the self-assembly
of regular radical clusters. Details of the structural and electronic
properties of BDPA clusters on Au(111) have been recently reported
by our group.[19] The STM topograph of Figure 1b shows a 15 × 15 nm2 frame of the
sample surface. Individual BDPA radicals are imaged as protrusions
with a characteristically curved “bean-like” shape.
Single isolated BDPA monomers are rarely observed on the sample surface
due to an increased surface mobility. Dimers are more stable due to
pinning at surface defect sites[19] and constitute
the smallest BDPA clusters. Larger clusters are 1-D chains up to several
nanometers long that start to grow at trimers with a characteristic
three-fold symmetric arrangement acting as nucleation centers for
the chain growth. (See Figure 1b.)
Evidence of
Kondo Behavior
The resonance energies of
BDPA’s frontier molecular orbitals (MOs) have been investigated
in detail by our group in ref (19). Figure 1c shows as an example the
dI/dV map at the energy of the highest
occupied molecular orbital (HOMO). Here we focus on the energy range
of ±100 meV close to EF, where Kondo
features are typically expected. At −100 meV, BDPA is almost
invisible in the dI/dV map against
the conductance background of the Au(111) surface,[19] where only the standing-wave pattern of the 2-D surface
state of Au(111) is observed (Figure 1d). In
addition to MO resonances, typical dI/dV spectra of BDPA/Au(111) exhibit a narrow conductance minimum (dip)
close to EF, whose amplitude is significantly
larger than the modulation (“ripples”) of the surface-state
band by the herringbone reconstruction of Au(111).[21] Figure 2a shows this conductance
minimum recorded with the STM tip over a single isolated BDPA monomer
(red curve) and a BDPA dimer (black curve). The tip position is marked
by a circle (○) in the insets showing the respective STM topographic
images. The maximum amplitude of the conductance dip is observed whenever
the STM tip is near the rim of BDPA. Amplitudes of up to 20% of the
zero-bias conductance of the pristine substrate are observed with
our best STM tips (Figure 2a). Over intramolecular
positions the conductance dip decreases (Figure 2b). A detailed analysis of our STS results on the conductance minimum
confirms a number of characteristic properties of the transmission
Kondo effect[11−14,22] for BDPA/Au(111):
Figure 2
STS results
of BDPA on Au(111). (a) dI/dV spectrum
of BDPA monomer (red) and a dimer (black) recorded
at 5 K over positions marked by circles in the STM topographic insets;
the feedback loop has been opened at +1 V, 50 pA, and a z-offset of 70 pm; the zero-bias conductance of pristine Au(111) was
calibrated to 1.3 nS; solid lines are numerical Fano fits of the Kondo
antiresonance (fit parameters see Table 1);
arrows mark broad shoulders attributed to inelastic excitations (details
see text); and vertical bars indicate energies of possible collective
vibrational modes obtained from DFT calculations relative to the Kondo
dip of the dimer marked by the dashed line. (b) dI/dV Kondo spectra over different positions (○)
of a BDPA dimer; scale bar = 0.2 nm. (c) dI/dV Kondo spectra of the BDPA dimer at different temperatures;
the feedback loop has been opened at +1 V, 50 pA, and a z-offset of 70 pm. (d) Temperature-dependent broadening of the Kondo
signal for the BDPA dimer recorded over similar STM tip position close
to the rim of BDPA; the dashed line represents best-fit based on the
formalism proposed by Nagaoka et al.[22]
(i) Energetic
position: The conductance minimum is centered at an energy offset,
|ΔE| = E – EF, of a few millielectronvolts around EF (Figure 2a). The width
of the signal is larger than the offset, 2Γ > ΔE, so that it spans over both the filled and empty state
regimes, that is, below and above EF.
According to Fermi-liquid theory, the value of ΔE is related to the electron–hole asymmetry.[17] The asymmetry is determined by the alignment of the MO
level that is singly occupied (SOMO) relative to EF and the Coulomb charging energy U that
separates SOMO from the singly unoccupied level (SUMO). For BDPA/Au(111),
the SOMO was found to lie several hundred millielectronvolts below EF,[19] and BDPA exhibits
a large U on the order of 1.4 eV.[23] The large U value is responsible for the
high stability of the radical state and strongly reduces electron
tunneling for the SOMO/SUMO levels in STM experiments (consistent
with the small q value obtained from the Fano fits
below). This hampers an accurate determination the electron–hole
asymmetry by tunneling spectroscopy. We observe ΔE < 0 for monomers and ΔE > 0 for dimers
(Table 1), suggesting that ΔE is affected by the presence of neighboring BDPA radicals.
Table 1
Fano Parameters of
BDPA/Au(111)a
q (±0.1)
ΔE (±0.7 meV)
Γ (±1 meV)
monomer
–0.15
–5.9
11
dimer
–0.23
+2.3
13
Fano fit parameters
of the BDPA
monomer (average of four) and dimer (average of three) adsorbed on
Au(111) at 5 K; asymmetry parameter q, energy offset
ΔE, and half-width Γ.
(ii) Spectral shape: The spectral shape of the conductance minimum
of Figure 2a is well-fitted by the Fano function[24] (compare experimental spectra with numerical
fits shown as black and red solid lines in Figure 2a). The values of the fit parameters are listed in Table 1. In the Fano picture, the shape of the resonance
is described by the asymmetry parameter, q. The numerical
value of q depends on the quantum mechanical phase
difference between different coherent spin-conserving tunneling paths
of electrons between STM tip and Kondo system[25,26] and determines whether a dip (reduced conductance) or peak (increased
conductance) is observed. For BDPA/Au(111), we obtain an almost constant
value of q ≈ −0.2 for monomers and
dimers. (See Table 1.) The small absolute value
of q indicates that electrons tunnel predominantly
between the tip and the continuum of substrate states rather than
the molecular impurity state.(iii) Temperature dependence:
Their lower surface mobility compared
with monomers facilitates the recording of dI/dV spectra of BDPA dimers up to 50 K (Figure 2c). With increasing temperature, the signal width, 2Γ,
increases and the amplitude decreases simultaneously, as expected
for surface Kondo systems.[11,12] Numerical values of
Γ determined by Fano fits of each spectrum are plotted against
the substrate temperature in Figure 2d for
the BDPA dimer. We have further analyzed the temperature broadening
based on a Fermi liquid description put forward by Nagaoka et al.[22] The least-squares fit based on this formalism,
Γ = 2((πkBT)2 + 2(kBTK)2)1/2, is shown as a dashed curve
in Figure 2d and yields a nominal value of TK = 54 ± 5 K for the BDPA dimer. This value
is consistent with the requirement of Fermi liquid theory for the
existence of Kondo physics[17] that the width
of the resonance, 2Γ, has to be much smaller than the spectral
energy of the bare impurity level (see above).(iv) Vibrational
satellites: The conductance minimum is accompanied
by two broad shoulders symmetrically offset above and below the Kondo
dip (marked by arrows in Figure 2a). The shoulders
indicate inelastic excitation of molecular vibrations by the tunneling
electrons. This is common for molecule-based Kondo systems, where
the impurity orbital is spatially extended and thus sensitive to geometric
changes of the molecular backbone.[13,27] The shoulders
in Figure 2a range from ∼25 to 90 meV
above and below the Kondo dip. This is significantly larger than the
expected thermal broadening of a single vibrational mode and points
to the excitation of a number of different vibrational modes with
wave numbers between ∼200 and 750 cm–1. Most
likely, collective backbone vibrations of BDPA are involved because
plain C–H vibrations typically exhibit wave numbers higher
than 1000 cm–1. This interpretation is supported
by density functional theory calculations of BDPA in the gas phase
performed previously by our group.[19] The
vertical bars in the top part of Figure 2a
mark the calculated energies of collective backbone vibrations; they
are symmetrically aligned above and below the Kondo minimum of the
dimer (marked by dashed line). Indeed, a number of calculated energies
lie within the broad symmetric shoulders of the spectrum.STS results
of BDPA on Au(111). (a) dI/dV spectrum
of BDPA monomer (red) and a dimer (black) recorded
at 5 K over positions marked by circles in the STM topographic insets;
the feedback loop has been opened at +1 V, 50 pA, and a z-offset of 70 pm; the zero-bias conductance of pristine Au(111) was
calibrated to 1.3 nS; solid lines are numerical Fano fits of the Kondo
antiresonance (fit parameters see Table 1);
arrows mark broad shoulders attributed to inelastic excitations (details
see text); and vertical bars indicate energies of possible collective
vibrational modes obtained from DFT calculations relative to the Kondo
dip of the dimer marked by the dashed line. (b) dI/dV Kondo spectra over different positions (○)
of a BDPA dimer; scale bar = 0.2 nm. (c) dI/dV Kondo spectra of the BDPA dimer at different temperatures;
the feedback loop has been opened at +1 V, 50 pA, and a z-offset of 70 pm. (d) Temperature-dependent broadening of the Kondo
signal for the BDPA dimer recorded over similar STM tip position close
to the rim of BDPA; the dashed line represents best-fit based on the
formalism proposed by Nagaoka et al.[22]Fano fit parameters
of the BDPA
monomer (average of four) and dimer (average of three) adsorbed on
Au(111) at 5 K; asymmetry parameter q, energy offset
ΔE, and half-width Γ.In Figure 2b, the inelastic excitations
are not as clearly distinguishable from the Kondo signal as in Figure 2a, which hampers the accurate determination of the
Kondo parameters with respect to the precise location over the BDPA
molecule. We attribute the spectral deviations to different tip structures,
varying adsorption sites, excitation of lateral motion (diffusion)
and noise in the dI/dV signal. Nevertheless,
Figure 2b evidences considerable variations
of the Kondo amplitude over different positions of the radical. The
Kondo amplitude of BDPA reaches maximum values close to the rim compared
with decreased values over the radical, similar to metal–organic
adsorbates with significant spin contributions from the hydrocarbon
backbone.[28,29] The spatial distribution of the Kondo amplitude
over the molecule is commonly assigned to the distribution of the
unpaired spin in the impurity orbital.[29] In that respect, Figure 2b is consistent
with the SOMO of BDPA obtained by calculation.[19]
Spin and Charge States of BDPA/Au(111)
The existence
of a Kondo signal indicates that the BDPA radicals remain magnetic
after adsorption on Au(111). Our recent electron spin resonance (ESR)
experiments on samples with BDPA monolayer coverage on Au(111)/mica
substrates have yielded g = 1.96 at 7 K.[19] This value is in good agreement with the isotropic g factor of g = 2.008 of the BDPA crystalline
bulk phase at 4 K and g = 2.0026 at room temperature.[30] The observed ESR activity of monolayer BDPA/Au(111)
together with the small g shift compared with bulk
phase suggest that a possible charge transfer from the Au substrate
to the adsorbed BDPA is small. This is corroborated by the standing
wave pattern of the Au(111) surface state shown in Figure 1d, where we find no evidence of an enhanced electrostatic
scattering potential at the BDPA radicals that would indicate a significant
charge transfer.[31] The above results corroborate
that the BDPA radicals preserve the initial radical spin-1/2 state
upon adsorption on the Au(111) surface.
Surface-State Contribution
to Kondo Screening
Figure 3 illustrates
the results of distance-dependent STS
measurements along a specific high-symmetry direction of a BDPA dimer.
STS spectra were recorded with the STM tip over positions at increasing
lateral distance, r, from the center of BDPA, as
marked by the circles in Figure 3a. The respective
spectra shown in Figure 3b exhibit the characteristic Kondo dip even over the
pristine Au(111) surface at large lateral distance of up to 1.5 nm
from the center of the radical. Outside of BDPA the signal amplitude
decreases monotonically according to 1/r. (See detailed
analysis in Figure 3c.) The 1/r dependence suggests that 2-D electronic states contribute to Kondo
screening.[26,32] In the present case, this points
to the surface-state band of Au(111), which acts as a 2-D electron
gas.
Figure 3
Decay of the Kondo signal
with lateral distance. (a) STM topography
of BDPA dimer recorded at +1 V and 50 pA; scale bar = 1 nm. (b) dI/dV Kondo spectra recorded at different
tip positions marked by the circles in panel a; the feedback loop
has been opened at +1 V, 50 pA, and a z-offset of
70 pm. (c) Decay of the Kondo amplitude with increasing lateral distance, r, from the center of BDPA; dashed line is a fit of 1/r.
Our experimental results indicate that the two BDPAs of
a dimer are individual spin-1/2 Kondo systems. Nevertheless, their
spins appear to be correlated through the surface-state: The two Kondo
screening clouds are so large that they may readily overlap; that
is, the two BDPAs of the dimer share part of a common Kondo cloud.
This seems to stabilize the individual screened states, as evidenced
by the higher TK of dimers compared with
monomers. (See Table 1.)Decay of the Kondo signal
with lateral distance. (a) STM topography
of BDPA dimer recorded at +1 V and 50 pA; scale bar = 1 nm. (b) dI/dV Kondo spectra recorded at different
tip positions marked by the circles in panel a; the feedback loop
has been opened at +1 V, 50 pA, and a z-offset of
70 pm. (c) Decay of the Kondo amplitude with increasing lateral distance, r, from the center of BDPA; dashed line is a fit of 1/r.
Conclusions
In
conclusion, our experimental STM and STS results obtained at
submolecular resolution from monomers and dimers of stable hydrocarbon
π radicals on Au(111) provide conclusive evidence of establishing
BDPA/Au(111) as a novel Kondo system, where a radical state in gas-phase
is preserved after adsorption on a metal surface.
Authors: Stefan Müllegger; Mohammad Rashidi; Michael Fattinger; Reinhold Koch Journal: J Phys Chem C Nanomater Interfaces Date: 2012-10-04 Impact factor: 4.126
Authors: Ryan Requist; Silvio Modesti; Pier Paolo Baruselli; Alexander Smogunov; Michele Fabrizio; Erio Tosatti Journal: Proc Natl Acad Sci U S A Date: 2013-12-23 Impact factor: 11.205