Tuning Overbias Plasmon Energy and Intensity in Molecular Plasmonic Tunneling Junctions by Atomic Polarizability

Plasmon excitation in molecular tunnel junctions is interesting because the plasmonic properties of the device can be, in principle, controlled by varying the chemical structure of the molecules. The plasmon energy of the excited plasmons usually follows the quantum cutoff law, but frequently overbias plasmon energy has been observed, which can be explained by quantum shot noise, multielectron processes, or hot carrier models. So far, clear correlations between molecular structure and the plasmon energy have not been reported. Here, we introduce halogenated molecules (HS(CH2)12X, with X = H, F, Cl, Br, or I) with polarizable terminal atoms as the tunnel barriers and demonstrate molecular control over both the excited plasmon intensity and energy for a given applied voltage. As the polarizability of the terminal atom increases, the tunnel barrier height decreases, resulting in an increase in the tunneling current and the plasmon intensity without changing the tunneling barrier width. We also show that the plasmon energy is controlled by the electrostatic potential drop at the molecule–electrode interface, which depends on the polarizability of the terminal atom and the metal electrode material (Ag, Au, or Pt). Our results give new insights in the relation between molecular structure, electronic structure of the molecular junction, and the plasmonic properties which are important for the development of molecular scale plasmonic-electronic devices.


■ INTRODUCTION
−5 Plasmons at interfaces can be excited not only by photons but also by high energy electrons in transmission electron microscopy (TEM) 6−8 or low energy tunneling electrons in scanning tunneling microscopy 9−11 and metal−insulator−metal (MIM) tunnel junctions 12−24 via inelastic tunneling (IET) which may find applications as nanoscale light sources, data processing devices, or integrated optical circuits.−31 We have shown that the generated plasmons can be controlled at the molecular level, 16 where orientation of the molecules modulates polarization, 32 and the structure of the molecules determines intensity and the dynamics 33 of the light emission.However, molecular plasmonic tunneling junctions are still not clearly understood for several reasons including: (i) the origin of over bias emission is still unclear, 28,34−42 (ii) the shape of the tunneling barrier is usually assumed to be rectangular (following the Simmons model) which is probably not justified (especially when dipoles or polarizable moieties 43−45 are present), and (iii) large voltages 46,47 (≫3 V) are applied which can lead to photon emission unrelated to excitation by IET (such as heating 39,40 ).Here we show, by simply changing one terminal atom in the molecular chemical structure, that the polarizability α of the terminal atom modifies the tunnel barrier shape and electrostatic potential profile in the molecular junction, which then affects both plasmon intensity (by 2 orders of magnitude) and overbias plasmon emission energy (over a range of 0.2 eV) emitted from these junctions without changing the barrier width at low V of ≤2.4 V.
3][14][15]18,19,23,24 From an electronics point of view, both the tunnel barrier width d and tunnel barrier height φ affect the electrical properties of MIM junctions, as they determine the tunneling rate across the junction and hence the plasmon intensity. 48−50 Althogh how the shape of the junction affects the plasmonic properties has been frequently investigated, 48 it is unclear to which extent the barrier properties affect the plasmonic properties.Often the shape of the tunneling barrier is modeled as a simple rectangular barrier using the Simmons model (see Background in the Supporting Information), but this kind of analysis is too simplistic to ignore; for instance, the presence of interface dipoles, 51,52 image charges, 53−55 and Fermi level pinning 56,57 all affect the shape of the tunneling barrier.In molecular junctions, the shape of the tunneling barrier can be quite complex depending on the presence of redox-active groups, 58 molecular dipoles, 51,52 or polarizable groups, 43−45 all of which can screen the electric field and thus alter the energy that is available for the tunneling charge carriers to excite plasmons.
All of these considerations complicate our understanding of the relation between d and φ and the energy of the plasmons.More specifically, for plasmon excitation by tunneling electrons, the energy of the photons (hv) that escapes the junction should not be higher than the energy provided by the bias V across the junction following the quantum cutoff law 12 given by eq 1 hv eV (1)   where h is Planck's constant, v is the frequency of the photon, and e is the elementary charge.Indeed, most studies report cutoff photon energies E c ≈ eV, 9−24 but exceptions with E c > eV 28,34−42 or E c < eV 46,47 have also been observed.The underlying mechanisms that cause these deviations are not fully disclosed.−42 McCreery and co-workers attributed the subthreshold light emission (E c < eV) to inelastic energy loss of the charge carriers during charge transport along the molecular chain although the mechanism of the light emission in their experiments is unclear considering the high applied bias up to 9 V. 46,47 Here we report molecular electronic plasmon sources based on self-assembled monolayer (SAM) tunnel junctions (STJs) where the shape of the tunneling barrier and electrostatic potential profile (see the Background in the Supporting Information) are determined by the properties of the SAM and the metal electrodes.We systematically changed α of the terminal group of the SAM, thereby changing the tunnel barrier shape and the effective potential drop along the SAM.We show that the plasmon energy (or E c ) can be related to the electrostatic potential profile of the molecular tunneling barrier φ, which includes effective potential drop along the SAM, and the built-in fields V int that are always present at metal− insulator interfaces due to the interface dipoles. 51,52To study the role of changes in work functions and Fermi-level pinning, we changed the material of the bottom electrode and found that E c changes with the work function shift ΔΦ (the shift of metal work function after modification of SAM, Pt > Au > Ag).Our results show that E c directly relates to the potential drop and electrostatic potential profile of the junctions, leading to overbias emission of up to 0.2 eV, which also gives new insights into the plasmon excitation mechanism.

■ RESULTS AND DISCUSSION
Experimental Design of the Plasmonic Junctions.Figure 1a shows a schematic of the STJs.We used templatestripped bottom-electrodes of Ag, Au, or Pt to support the SAMs of S(CH 2 ) 12 X (in short SC 12 X) where X = H, F, Cl, Br, or I and formed electrical top contacts to these SAMs of EGaIn with the well-established EGaIn technique 59 (EGaIn = eutectic gallium indium alloy).Although Pt is a more lossy plasmonic material than Au or Ag, 60,61 plasmons on Pt have been excited by both optical 62 and electrical 63 means.Besides, junctions based on Pt bottom electrodes have a higher breakdown voltage than the Au and Ag counterparts 64 which facilitate the optical measurements in the high bias regime.Therefore, we focus here on Pt substrate where we systematically change the type of the atom of the headgroup X (to ensure to keep supramolecular changes to the monolayer structure negligible).Figure 1b shows the corresponding energy level diagram of these SC 12 X junctions on Pt where the shape of the tunnel barrier is indicated by the dashed lines.With this series of SAMs, d is kept similar, while φ decreases from F to I by ∼5 eV 43 as α of SC 12 X increases as shown in Figure 1b (see below for details).Polarizable terminal atoms screen electric fields more efficiently than nonpolarizable ones as α directly relates to the dielectric constant (as, for instance, given by the Clausius-Mossotti relation 65 ), which, in turn, affects the ionization potential (or electron affinity).We have shown elsewhere that indeed the relative dielectric constant ε r of these junctions increases from 2.5 to 8.9 as X goes from F to I, enhancing the tunneling rates by 4 orders of magnitude, thus effectively reducing the effective tunneling barrier height. 43We hypothesize that, as the electric-field screening capability of the halogens increases, the potential drop at the SAM//EGaIn interface increases (Figure 1c) which effectively reduces the voltage, V eff , leading to a reduced excited plasmon energy (see below).
On the other hand, with the same SAM, we also changed the material of the bottom electrode and thus changed the interface energetics (see Background in the Supporting Information) at the SAM-bottom electrode interface.As shown in Figure 1d, although the native metals have different work functions Φ M , after SAM formation, the modified work functions Φ M-SAM are similar due to Fermi-level pinning. 56,57his Fermi-level pinning results in a work function shift ΔΦ (Pt > Au > Ag) and associated formation of interface dipoles.These molecular interface dipoles are indicated by the black arrows.We hypothesize that V int will also affect E c (see below).
Characterization of the SAMs.The SAM precursors HSC 12 X and the corresponding SAMs on Au, Ag, and Pt were prepared using previously reported methods (see Method section in the Supporting Information for details). 43−68 The SAMs derived from HSC 14 X on Ag have also been reported. 43Below we have confirmed by X-ray photoelectron spectroscopy (XPS) that all SAMs have a similar surface coverage (Γ SAM , in nmol•cm −2 ) on Pt (Table 1 and the Method in the Supporting Information).In agreement with previous reports, 43,45 we conclude that our SAMs formed wellorganized and densely packed structures.
With a similar HOMO energy, the LUMO energy decreases from X = F to X = I, which verifies the tunnel barrier shape, as shown in Figure 1b (and Supporting Information Figure S10).In other words, the tunneling barrier height (which is defined as the offset in energy between the LUMO and Fermi-level of the electrode) decreases from X = F to X = I.On the other hand, previous impedance measurements have shown increased ε r by 3.5 times from X = F to X = I, 43 as α is directly related to ε r according to the Clausius-Mossotti relation. 65Thus, polarizable terminal atoms (e.g., X = I) screen electric fields more efficiently than nonpolarizable atoms (e.g., X = F), which changes the electrostatic potential profile as sketched in Figure 1c with larger potential drop for X = I than that for X = F at the monolayers−top electrode interface.
Shift in Work Function and Interface Dipoles.When molecules are brought into contact with the metal surface, metal−molecule bonds form (see the Background in the Supporting Information), leading to changes in the energylevel alignment and formation of interface dipoles.With ultraviolet photoelectron spectroscopy (UPS; see Method section in the Supporting Information), we determined the Φ M-SAM (eV) using well-established methods. 43The work function change ΔΦ of the metal surfaces induced by the modification of SAMs can be calculated from the difference between Φ M-SAM and Φ M , which directly reflects the electrostatic potential steps at the SAM-bottom electrode interface (as drawn in Figure 1d).The interface dipole is determined by both μ bond (the dipole of metal−sulfur M−S bond) and μ mol,⊥ (the dipole of molecule along the surface normal) given by the Helmholtz equation (eq 2) where e is the elementary charge (1.602 × 10 −19 C), N is the dipole density (here it corresponds to Γ SAM , in nmol•cm −2 ), μ bond and μ mol,⊥ are the bond and molecular dipole moments (along surface normal direction, in D), ε 0 is the vacuum permittivity (8.85 × 10 −12 F/m), and ε r is the relative dielectric constant.Since we varied the SAM and the bottom electrode separately, we can extract both the values of μ bond and μ mol,⊥ for each terminal group X (see Method section in the Supporting Information, Table S1 and Table S2).Figure 2 shows the plot of ΔΦ, μ bond , and μ mol,⊥ for SC 12 H SAM on Ag, Au, and Pt substrates and for SC 12 X (X = F, Cl, Br, and I) SAMs on Pt substrates (see Method in Supporting Information for more details of the UPS analysis).According to Figure 2a, a   In other words, the work function of all three SAM coated metals is essentially the same which is well-known and is caused by the push-back effect (or Fermi-level pinning). 56Further, due to the different nature of the M−S bond, the Au−S bond shows the smallest value of μ bond while Ag−S and Pt−S bonds show larger values of μ bond but with different signs (Figure 2c), indicating the opposite direction of the two bond dipoles. 70For SAMs with different terminal atoms, the extracted μ mol,⊥ values change from −0.5 to −2.27 D within experimental error (Figure 2d).
Electrical Characterizations of the STJs.To form the STJs, we used the EGaIn technique with the EGaIn electrode stabilized in a through-hole in a microfluidic network made from PDMS (polydimethylsiloxane) following a previously reported procedure 59 (see Methods in the Supporting Information).We characterized these halogenated junctions electrically by collecting 50−200 J(V) curves for each type of SAM. Figure 3 shows the log-average J(V) curves of the junctions with SAMs of SC 12 X on Pt substrates recorded in the bias range of ±1.8 V (see Methods in the Supporting Information for details).The value of J increases by 2 orders of magnitude when changing from X = F or H to X = I in agreement with previously reported results. 43,45This increase in J is expected since the value of φ decreases for junctions with SAMs from X = F to I.
Excited Plasmon Intensity.To study the dependency of excited plasmon on the shape of the molecular tunnel barrier, we characterized the light emission, i.e., radiative decay of surface plasmons excited in the STJs, in the far field using an inverted optical microscope equipped with an electron multiplying charge-coupled device (EMCCD) and a 100× oil immersion objective with numerical aperture NA = 1.49following previously reported procedures. 16Knowing that the tunneling current increases for junctions with SAMs from X = F to I (Figure 3), we measured the light emission images from the area of the STJ with an integration time of 2 min for junctions with low currents with X = F and H or 30 s (to avoid the saturation of the detector) for the other three conductive junctions with X = Cl, Br, or I. Figure 4a−e shows the representative real plane emission images recorded at −1.8 V.The light emission originates from discrete spots due to the high surface roughness of the top electrode, resulting in a much lower effective electrical contact area than the geometrical footprint of the top electrode.Since tunneling currents decay exponentially with distance, light emission is only observed in the effective electrical contact regions (see ref 66 for details).These images also show that the photon emission intensity increases when X changes from F or H to I.This follows the same trend of the above-mentioned increasing J and decreasing φ, induced by increasing α of the same series of X.In these STJs, the HOMO orbital is pinned to the bottom electrode; thus, φ (Table 1) can be calculated as the difference between the measured Φ M-SAM (eV) and the calculated LUMO energy, and should have a linear relationship with log 10 |J| and hence log 10 |I| based on eq S1 (see Background in Supporting Information), as shown in Figure 4f.
Here, the photon emission density I (in counts/s/cm 2 ) is defined as the total photon counts divided by the integration time and geometric contact area of EGaIn.The value of J and   photon emission density I (in counts/s/cm 2 ) for each SAM was determined at −1.8 V (see Methods in the Supporting Information).Figure 4f shows that both log 10 |J| and log 10 |I| decrease linearly as increases, showing a similar slope.It thus indicates that there is also a linear correlation between the tunneling current and the photon emission intensity, which agrees with our previous report. 16xcited Plasmon Energy: Atomic Polarizability Dependency.To measure the excited plasmon energy, we also recorded the light emission spectra to determine how E c changes as a function of X (see Methods in the Supporting Information).Figure 5a shows the representative spectra of the emitted photons at an applied bias of −2.0 V and that the spectra red shift from F or H to I by ∼0.2 eV.We note that our junctions consist of planar electrodes (and not of plasmonic resonators or antennas 11,15,18 ); therefore, the red shift is not related to the electrode geometry, but it is a molecular effect.This observation confirms our hypothesis that emitted plasmon energy depends on the shape of the molecular tunnel barrier and the resulting electrostatic potential profile across the junction (see Figure 1c) since the other parameters of the tunneling barrier were left unchanged.
By recording E c as a function of applied bias (Figure 5b), we observe an over bias emission (hv > eV) across the measured bias range for all five molecules, which we explain as follows.As the same metal electrodes (Pt and EGaIn) were used for the five molecules, the red-shift of the plasmon spectra in Figure 5a,b should be a molecular effect arising from the difference in X, i.e., the SAM//EGaIn interface.To correlate this red-shift to the electrostatic potential drop at the interface induced by α of the terminal atom, we determined the reduced photon energy ΔE (%) relative to X = H with eq 3: where ΔE is defined as the percentage of the change in photon energy at each voltage.Table 2 summarizes E c and ΔE at each voltage and shows ΔE increases in the sequence: I > Br > Cl > F ≈ H. Further, ΔE was plotted against ε r (Figure 5c,d) which shows a positive correlation.As ε r is directly related to α, these results prove that α can induce a red-shift in the emission photon energy by modifying the electrostatic potential profile of the STJ.The decrease of overbias emission going from X = F to X = I can be explained as the interface dipoles play a significant role for X = F but are screened for X = I.The polarizability α indicates the ability of an atom or molecule to respond to changes in an electric field by forming induced dipoles.More specifically, the electron clouds of large atoms are distorted more in an external electric field than in small atoms, leading to a dipole.In our work, as X goes from F to I, the value of α increases leading to increasingly large dipoles in applied electric fields.This screening of the electric field by increasingly large halogen atoms leads to an increasingly large potential drop at the SAM//EGaIn interface, which effectively reduces the voltage, V eff , and decreases the excited plasmon energy.In other words, for monolayers with X = F, the aligned interfaces dipoles with the electric field of the applied bias results in a larger plasmon excitation energy than expected from the applied bias alone leading to overbias emission.In contrast, for monolayers with X = I that can screen, the effect of dipoles will lead to emission close to the quantum cutoff values.Below we carry out additional experiments to test this hypothesis in more detail by systematically varying μ bond .
In both Figure 5 and Table 2, we observed that E c > eV.To make sure we are far away from the quantum point contact regime which could result in above quantum threshold photon emission, an estimation was carried out below.Using SC 12 H SAM as an example, the current of the STJ at −1.8 V is ∼100 nA (Figure 3), corresponding to junction conductance G J of 7.2 × 10 −4 G 0 .By further considering the number of light emission spots of 30 (Figure 4, V = −1.8V) in the junction, the conductance under each light emitting spot is . In case of the SC 12 I SAM, G S increases by 2 orders of magnitude, and that is 10 −3 G 0 .Based on the above estimation, we can safely conclude that the models involving electron−electron interactions (in the quantum conductance regime) do not apply to our STJs.The model of Fermi level distribution can also be ruled out, as it cannot explain the substrate and bias polarity effect we observed below (Figure 6).Excited Plasmon Energy: Interface Dipole Dependency. Figure 2 shows a large difference in ΔΦ among different metal substrates, induced by the different Φ M values and hence the different interface dipoles.As mentioned above, the offset between E c and eV in our STJs should originate from the dipoles at the molecule−electrode interface, which add to or reduce the effective electric field under externally applied bias.To determine how the μ bond component of the interface dipole affects the excited plasmon energy, we recorded the light emission spectra of SC 12 H junctions on Ag, Au, and Pt substrates at −2.0 V bias. Figure 6a shows that the spectra recorded from the Ag and Au junctions show E c ≈ eV, while that recorded from Pt gives an ∼0.2 eV blue-shift resulting in E c > eV.The offset between E c and eV qualitatively matches the value of ΔΦ for the three metals (Figure 2) and proves that indeed the interface dipoles and associated built-in potentials at the molecule−electrode interface have an influence on the excited plasmon energy.Since ΔΦ is independent of the applied bias, upon reversal of the bias, we expect a red shift.Figure 6b shows, indeed, an ∼0.1 eV red-shift on Pt at positive bias polarity and now E c < eV applies.The offsets between E c and eV are much smaller than the ΔΦ measured by UPS which may be caused by the presence of other interface dipoles at, for instance, the SAM//EGaIn interface 71 or rounding of the tunneling barriers (as mentioned before, assuming rectangular tunneling barriers is like being too simplistic).Note that we increased the electrical bias to ±2.3 V to record the emission spectra shown in Figure 6b, because the light emission was too weak at smaller applied bias in the positive direction (see Methods in the Supporting Information).
■ CONCLUSIONS Atomic Polarizability Changes the Tunnel Barrier and Electrostatic Potential and Plasmonic Properties of Tunnel Junctions.This atomic control originates from the screening of the electric field in the junction, which, in turn, affects the excited plasmon or photon emission energy: molecules with higher α screen the electric field more efficiently, resulting in a lower tunnel barrier, larger potential drop, and thus lower photon energy, and vice versa for molecules with lower α.While in our case, even though there is no hopping (based on the molecular structure of our SAMs and the absence of a thermally activated component in the charge transport process 43,45 ), we have evidence for above and below threshold photon emission depending on the terminal atom and work function (and associated interface dipole; Figure 1d).Since interfacial dipoles are directional, we found above or below threshold emission, depending on the applied bias polarity.−42 Probing Interface Energetics with Excited Plasmon Energy.We also show that by measuring the excited plasmon energy in these junctions as a function of the metal substrates the interface energetics (at the molecule-bottom electrode interface) can be probed qualitatively.The energy of the excited plasmon or emitted photon depends on the effective potential drop across the junction.The presence of interface dipoles (and Fermi-level pinning) generates built-in electric fields that are aligned with, or against, the externally applied electric field and thus increase or lower the energy available to excite plasmons.Experimentally, we show that, with three different metals, the observed photon energy offset (between E c and quantum cutoff) qualitatively match with the work function shift (corresponding to the interface dipoles and builtin electric fields) measured with UPS.Moreover, with the same metal, the change of sign in photon energy offset upon reversing the bias polarity further confirms the role of the builtin fields.Here, we can only be qualitative as UPS only characterized the single interface, i.e., molecule-bottom electrode interface, while we know that, once the junction form, there is another molecule-top electrode interface, i.e., SAM//EGaIn interface; as a result, all the energy levels will realign, which will change the interface energetics.On the other hand, the observed lower photon energy offset compared to work function shift measured by UPS may also be related to the formation of image charges 53−55 in the metal electrode.For example, in EGaIn junctions, image charge effect leads to a modest (1.5 times) reduction in the HOMO−LUMO gap and associated tunneling barrier heights, 72 but in single-molecule junctions, image charge effects may be 2−3 times larger. 53uch effects could be studied in the future by potentiodynamic approaches, for example, where the spectra of the light emission are systematically studied as a function of applied voltage.

New Insights Regarding above and below
A Plasmonic Approach to Characterize Molecular Junctions in Operando.Although most research on plasmonic tunnel junctions has optical applications in mind, our results suggest that the plasmonic properties can reveal information on molecular tunneling junctions difficult, or perhaps impossible, to obtain with other techniques, such as potential drop profiles or the effect of interface dipoles on tunneling electronics by examining shifts in photon energies.Therefore, we believe that our findings appeal to both the molecular electronics and plasmonics communities and that our results will stimulate new research directions.

Figure 1 .
Figure 1.(a) Schematic representation of the STJs with SC 12 X (X = H, F, Cl, Br, or I) SAMs on different metal substrates (M = Ag, Au, or Pt).(b) Energy level diagram of the Pt-SC 12 X//EGaIn junctions with zero bias, where X is hydrogen or halogen, as stated in panel a.The shapes of the tunnel barriers are indicated by the dashed lines which change as a function of X. (c) The shape of the electrostatic potential profiles across the Pt-SC 12 X//EGaIn junctions (EGaIn stands for eutectic alloy of gallium and indium) with V Bias to the EGaIn top electrode (in all of our experiments, the bottom electrode was grounded).(d) Interface energetics of the SC 12 -bottom electrode interface.Φ M and Φ M-SAM are work functions of the bottom electrodes before and after SAM formation.The black arrows indicate the relative magnitude of the molecular interface dipoles.
b HOMO was obtained by adding UPS measured Φ Pt-SAM and HOMO onset.Systematic error is ±0.05 eV.c LUMO was obtained by adding E g (taken from ref 43) to the HOMO energy.Systematic error is ±0.05 eV.d φ was calculated as the energy offset between Φ Pt-SAM and LUMO.e α is the atomic polarizability for the terminal halogen atoms taken from ref 69.−1.03 eV when M changes from Ag to Pt.This change in ΔΦ offsets the overall change in the work function.

Figure 2 .
Figure 2. (a) Plot of ΔΦ for SC 12 H SAM on Ag, Au, and Pt.(b) Plot of ΔΦ for SC 12 X (X = F, Cl, Br, and I) SAMs on Pt.(c) Plot of μ bond for SC 12 H SAM on Ag, Au, and Pt.(d) Plot of μ mol,⊥ for SC 12 X (X = F, Cl, Br, and I) SAMs on Pt.

Figure 3 .
Figure 3. J(V) characterization of the STJs with SC 12 X SAMs at ±1.8 V on a Pt substrate.The error bars stand for the log-standard deviations.

Figure 4 .
Figure 4. Representative real plane emission images at −1.8 V with SC 12 H (a), SC 12 F (b), SC 12 Cl (c), SC 12 Br (d), and SC 12 I (e) SAMs.(f) log 10 |J| and log 10 |I| at −1.8 V plot as a function of .The dashed lines are linear fits to the data, confirming the dependency of the light emission intensity and tunneling current on the barrier height.

Figure 5 .
Figure 5. (a) Light emission spectra recorded at −2.0 V bias on Pt substrates with different SAMs.(b) Plot of E c as a function of voltage with SC 12 X STJs.The dashed line in b indicates the quantum cutoff with hv = eV.(c, d) The correlation between the reduced photon energy ΔE and ε r at different biases (c) and the average from the three biases (d), with error bars from the standard deviation.The lines are linear fits to data.These correlations confirm that α of the SAM terminal atoms changes the electrostatic potential profile of the junctions, tuning the excited plasmon energy.
Threshold Photon Emission.Compared to the literature examples introduced in the Background section (Supporting Informa-tion), our junctions are far below the quantum conductance and thus are closer to the molecular junctions investigated by McCreery et al.They claimed that the below threshold photon emission in their systems comes from the electron energy loss during charge hopping along the molecule.

Figure 6 .
Figure 6.(a) Light emission spectra recorded on different bottom electrodes (Au, Ag, and Pt) at −2.0 V bias with SC 12 H junctions.(b) Emission spectra of SC 12 H junction on Pt electrode collected at ±2.3 V.

Table 1 .
clear trend from different metals in the value of ΔΦ (varying with Φ M ) is observed.The value of ΔΦ increases from −0.17 to Properties of the SC 12 X SAMs on Pt Substrates SAM is the relative surface coverage obtained from relative intensities of the Pt 4f peak vs that of SC 12 Br (with a surface coverage of 1.1 nmol•cm −2 according to ref 43).
a Γ

Table 2 .
E c and ΔE with Different X