Abstract
Background: The transport through a quantumscale device may be uniquely characterized by its transmission eigenvalues τ_{n}. Recently, highly conductive singlemolecule junctions (SMJ) with multiple transport channels (i.e., several τ_{n} > 0) have been formed from benzene molecules between Pt electrodes. Transport through these multichannel SMJs is a probe of both the bonding properties at the lead–molecule interface and of the molecular symmetry.
Results: We use a manybody theory that properly describes the complementary wave–particle nature of the electron to investigate transport in an ensemble of Pt–benzene–Pt junctions. We utilize an effectivefield theory of interacting πelectrons to accurately model the electrostatic influence of the leads, and we develop an ab initio tunneling model to describe the details of the lead–molecule bonding over an ensemble of junction geometries. We also develop a simple decomposition of transmission eigenchannels into molecular resonances based on the isolated resonance approximation, which helps to illustrate the workings of our manybody theory, and facilitates unambiguous interpretation of transmission spectra.
Conclusion: We confirm that Pt–benzene–Pt junctions have two dominant transmission channels, with only a small contribution from a third channel with τ_{n} << 1. In addition, we demonstrate that the isolated resonance approximation is extremely accurate and determine that transport occurs predominantly via the HOMO orbital in Pt–benzene–Pt junctions. Finally, we show that the transport occurs in a lead–molecule coupling regime where the charge carriers are both particlelike and wavelike simultaneously, requiring a manybody description.
Introduction
The number of transmission channels for a singleatom contact between two metallic electrodes is simply given by the chemical valence of the atom [1]. Recently, it was argued [2] that the number of dominant transmission channels in a singlemolecule junction (SMJ) is determined by the degeneracy of the molecular orbital [3] closest to the metal Fermi level. In this article, we focus on ensembles of highly conductive Pt–benzene–Pt junctions [4] in which the lead and molecule are in direct contact. Going beyond the phenomenological randommatrix model of lead–molecule coupling considered in [2], a realistic atomistic model is developed to describe lead–molecule coupling over an ensemble of energetically favored junction geometries.
For a twoterminal SMJ, the transmission eigenvalues τ_{n} are eigenvalues of the elastic transmission matrix [5]
where G is the retarded Green’s function [6] of the SMJ, Γ^{α} is the tunnelingwidth matrix describing the bonding of the molecule to lead α, and the total transmission function T(E) = Tr{T(E)}. The number of transmission channels is equal to the rank of the matrix (Equation 1), which is in turn limited by the ranks of the matrices G and Γ^{α} [2]. The additional twofold spin degeneracy of each resonance is considered to be implicit throughout this work. As indicated by Equation 1, an accurate description of transport requires an accurate result for G, which can be calculated by using either singleparticle or manybody methods, and which depends critically on accurate descriptions of the molecular energy levels and the lead–molecule coupling.
In effective singleparticle theories, including current implementations of density functional theory (DFT), it is often necessary [710] to describe the transport problem by considering an “extended molecule”, composed of the molecule and several electrode atoms. This procedure makes it difficult, if not impossible, to assign transmission eigenchannels to individual molecular resonances since the quantum states of the extended molecule bear little resemblance to the states of the molecule itself.
We utilize a nonequilibrium manybody theory based on the molecular Dyson equation (MDE) [6] to investigate transport distributions of SMJ ensembles. Our MDE theory correctly accounts for wave–particle duality of the charge carriers, simultaneously reproducing the key features of both the Coulomb blockade and coherenttransport regimes, alleviating the necessity of constructing an extended molecule. Consequently, we can unambiguously assign transmission eigenchannels to molecular resonances [2].
Previous applications of our MDE theory [6,11,12] to transport through SMJs utilized a semiempirical Hamiltonian [13] for the πelectrons, which accurately describes the gasphase spectra of conjugated organic molecules. Although this approach should be adequate to describe molecules weakly coupled to metal electrodes, e.g., by thiol linkages, in junctions where the πelectrons bind directly to the metal electrodes [4], the lead–molecule coupling may be so strong that the molecule itself is significantly altered, necessitating a more fundamental molecular model.
In this work, we utilize an effective field theory of interacting πelectrons (πEFT), in which the form of the molecular Hamiltonian is derived from symmetry principles and electromagnetic theory (multipole expansion) [14]. The resulting formalism constitutes a stateoftheart manybody theory that provides a realistic description of lead–molecule hybridization and van der Waals coupling, as well as the screening of intramolecular interactions by the metal electrodes, all of which are essential for a quantitative description of stronglycoupled SMJs [4].
The bonding between the tip of electrode α and the molecule is characterized by the tunnelingwidth matrix Γ^{α}, where the rank of Γ^{α} is equal to the number of covalent bonds formed between the two. For example, in a SMJ where a Au electrode bonds to an organic molecule through a thiol group, only a single bond is formed, and there is only one significant transmission channel [15,16]. In Pt–benzene–Pt junctions, however, each Pt electrode forms multiple bonds to the benzene molecule and multiple transmission channels are observed [4]. In such highly conductive SMJs the lead and molecule are in direct contact and the overlap between the πelectron system of the molecule and all of the quasiatomic wavefunctions of the atomically sharp electrode are relevant. Rather than the randommatrix method used in [2], we develop an atomistic approach to bonding in which the nine relevant orbitals for each Pt electrode are included (one s, three p, and five d), representing the evanescent tunneling modes in free space outside the apex atom of each electrode tip. This atomistic model of lead–molecule coupling allows distributions of transport coefficients to be computed directly over an ensemble of junction geometries, supplanting the phenomenological model of lead–molecule coupling employed in [2].
In the next section, we outline the relevant aspects of our MDE theory and derive transport equations in the isolatedresonance approximation. We then develop our atomistic treatment of lead–molecule coupling, in which the electrostatic influence of the leads is treated by πEFT and the multiorbital lead–molecule bonding is described using the quasiatomic orbitals of the electrode tip. Finally, the transport distributions for these ensembles of Pt–benzene–Pt junctions are calculated by using both the full molecular Green’s function and within the isolatedresonance approximation. The efficacy of the isolated resonance approximation is investigated in detail.
Manybody theory of transport
When macroscopic leads are bonded to a single molecule, a SMJ is formed, transforming the fewbody molecular problem into a full manybody problem. The bare molecular states are dressed by interactions with the lead electrons when the SMJ is formed, shifting and broadening them in accordance with the lead–molecule coupling.
Until recently [6] no theory of transport in SMJs was available which properly accounted for the particle and wave character of the electron, such that the Coulomb blockade and coherent transport regimes were considered “complementary” [10]. Here, we utilize a manybody MDE theory [6,12] based on nonequilibrium Green’s functions (NEGFs) to investigate transport in multichannel SMJs, which correctly accounts for both aspects of the charge carriers.
In order to calculate transport quantities of interest we must determine the retarded Green’s function G(E) of the junction, which may be written as
where H_{mol} = + is the molecular Hamiltonian, which we formally separate into onebody and twobody terms [6,12]. S is an overlap matrix, which in an orthonormal basis reduces to the identity matrix, and
is the selfenergy, including the effect of both a finite lead–molecule coupling, through , and manybody interactions, through the Coulomb selfenergy Σ_{C}(E). The tunneling selfenergy matrices are related to the tunnelingwidth matrices by
It is useful to define a molecular Green’s function In the sequential tunneling regime [6], where lead–molecule coherences can be neglected, the molecular Green’s function within MDE theory is given by
where all onebody terms are included in and the Coulomb selfenergy Σ^{(0)} accounts for the effect of all intramolecular manybody correlations exactly. The full Green’s function of the SMJ may then be found using the molecular Dyson equation [6]
where ΔΣ = Σ_{T} + ΔΣ_{C} and At room temperature and for small bias voltages, ΔΣ_{C} ≈ 0 in the cotunneling regime [6] (i.e., for nonresonant transport). Furthermore, the inelastic transmission probability is negligible compared to the elastic transmission in that limit.
The molecular Green’s function G_{mol} is found by exactly diagonalizing the molecular Hamiltonian, including all charge states and excited states of the molecule [6,12]
where is the probability that the molecular state ν is occupied, C(ν,ν′) are manybody matrix elements and In linear response, where is the grand canonical partition function.
The rank1 matrix C(ν,ν′) has elements
where d_{n}_{σ} annihilates an electron of spin σ on the nth atomic orbital of the molecule, and ν and ν′ label molecular eigenstates with different charge. The rank of C(ν,ν′) in conjunction with Equation 6 and Equation 7 implies that each molecular resonance ν → ν′ contributes at most one transmission channel in Equation 1, suggesting that an Mfolddegenerate molecular resonance could sustain a maximum of M transmission channels.
Isolatedresonance approximation
Owing to the position of the chemical potential of the leads relative to the molecular energy levels and the large charging energy of small molecules, transport in SMJs is typically dominated by individual molecular resonances. In this subsection, we calculate the Green’s function in the isolatedresonance approximation wherein only a single (nondegenerate or degenerate) molecular resonance is considered. In addition to developing intuition and gaining insight into the transport mechanisms in a SMJ, we also find (cf. Results and Discussion section) that the isolatedresonance approximation can be used to accurately predict the transport.
Nondegenerate molecular resonance
If we consider a single nondegenerate molecular resonance then
where ε = E_{ν′} − E_{ν}, is the rank1 manybody overlap matrix, and we have set In order to solve G analytically, it is useful to rewrite Dyson’s equation (Equation 6) as follows:
In the elasticcotunneling regime (ΔΣ_{C} = 0) we find
Equation 11 can be equivalently expressed as
where
is the effective selfenergy at the resonance, which includes the effect of manybody correlations through the C(ν,ν′) matrix.
Using Equation 1, the transmission in the isolatedresonance approximation is given by
where
is the dressed tunnelingwidth matrix, and
As evidenced by Equation 14, the isolatedresonance approximation gives an intuitive prediction for the transport. Specifically, the transmission function is a single Lorentzian resonance centered about with a halfwidth at halfmaximum of The lessintuitive manybody aspect of the transport problem is encapsulated in the effective tunnelingwidth matrices where the overlap of molecular manybody eigenstates can reduce the elements of these matrices and may strongly affect the predicted transport.
Degenerate molecular resonance
The generalization of the above results to the case of a degenerate molecular resonance is formally straightforward. For an Mfold degenerate molecular resonance
The M degenerate eigenvectors of G_{mol} may be chosen to diagonalize Σ_{T} on the degenerate subspace
and Dyson’s equation may be solved as before
Although Σ_{T} is diagonal in the basis of Γ^{L} and Γ^{R} need not be separately diagonal. Consequently, there is no general simple expression for T(E) for the case of a degenerate resonance, but T can still be computed using Equation 1.
In this article we focus on transport through Pt–benzene–Pt SMJs where the relevant molecular resonances (HOMO or LUMO) are doubly degenerate. Considering the HOMO resonance of benzene
where diagonalize Σ_{T} and 0_{N} is the Nparticle ground state.
πElectron effective field theory
In order to model the degrees of freedom most relevant for transport, we have utilized an effective field theory of interacting πelectron systems (πEFT) as described in detail in [14]. Briefly, this was done by starting with the full electronic Hamiltonian of a conjugated organic molecule and by dropping degrees of freedom far from the πelectron energy scale. The effective πorbitals were then assumed to possess azimuthal and inversion symmetry, and the effective Hamiltonian was required to satisfy particle–hole symmetry and be explicitly local. Such an effective field theory is preferable to semiempirical methods for applications in molecular junctions because the effective interaction is derived from Maxwell’s equations, and hence can be readily generalized to include screening of intramolecular Coulomb interactions due to nearby metallic electrodes.
Effective Hamiltonian
This allows the effective Hamiltonian for the πelectrons in gasphase benzene to be expressed as
where t is the tightbinding matrix element, μ is the molecular chemical potential, U_{nm} is the Coulomb interaction between the electrons on the nth and mth πorbitals, and The interaction matrix U_{nm} is calculated by way of a multipole expansion keeping terms up to the quadrupole–quadrupole interaction:
where U^{MM} is the monopole–monopole interaction, U^{QM} is the quadrupole–monopole interaction, and U^{QQ} is the quadrupole–quadrupole interaction. For two πorbitals with arbitrary quadrupole moments and and centers separated by a displacement , the expressions for these are
where
is a rank4 tensor that characterizes the interaction of two quadrupoles and is a dielectric constant included to account for the polarizability of the core and σ electrons. Here i, j, k, and l are the Cartesian indices of the foregoing tensors and vectors. Altogether, this provides an expression for the interaction energy that is correct up to fifth order in the interatomic distance.
Benzene
The adjustable parameters in our Hamiltonian for gasphase benzene are the nearestneighbor tightbinding matrix element t, the onsite repulsion U, the dielectric constant , and the πorbital quadrupole moment Q. These were renormalized by fitting to experimental values that should be accurately reproduced within a πelectron only model. In particular, we simultaneously optimized the theoretical predictions of (1) the six lowest singlet and triplet excitations of the neutral molecule, (2) the vertical ionization energy, and (3) the vertical electron affinity. The optimal parametrization for the πEFT was found to be t = 2.70 eV, U = 9.69 eV, Q = −0.65 eÅ^{2} and = 1.56 with a RMS relative error of 4.2 percent in the fit of the excitation spectrum. It would be interesting to compare the values of U and Q determined by this analysis with estimates from ab initio methods such as density functional theory. Note, however, that the use of “improved” values of the parameters in our effective Hamiltonian is unlikely to improve agreement with the experimental data that we considered, precisely because we optimized the πEFT predictions for these quantities.
The top panel of Figure 1 shows the spectral function for gasphase benzene within πEFT, along with experimental values for the first optical excitation of the cation (3.04 eV), the vertical ionization energy (9.23 eV), and the vertical electron affinity (−1.12 eV). As a guide for the eye, the spectrum has been broadened artificially by using a tunnelingwidth matrix of Γ_{nm} = (0.2 eV)δ_{nm}. The close agreement between the experimental values and the maxima of the spectral function suggests that our model is accurate at this energy scale. In particular, the accuracy of the theoretical value for the lowest optical excitation of the cation is noteworthy, as this quantity was not fit during the renormalization procedure but rather represents a prediction of πEFT.
In order to incorporate screening by metallic electrodes into πEFT, we utilized an image multipole method whereby the interaction between an orbital and image orbitals is included up to the quadrupole–quadrupole interaction in a screened interaction matrix In particular, we chose a symmetric that ensures the Hamiltonian gives the energy required to assemble the charge distribution from infinity with the electrodes maintained at fixed potential, namely
where U_{nm} is the unscreened interaction matrix and is the interaction between the nth orbital and the image of the mth orbital. When multiple electrodes are present, the image of an orbital in one electrode produces images in the others, resulting in an effect reminiscent of a hall of mirrors. We deal with this by including these “higher order” multipole moments iteratively until the difference between successive approximations of drops below a predetermined threshold.
In the particular case of the Pt–benzene–Pt junction ensemble described in the next section, the electrodes of each junction are modeled as perfect spherical conductors. An orbital with monopole moment q and quadrupole moment Q^{ij} located a distance r from the center of an electrode with radius R then induces an image distribution at with monopole and quadrupole moments
and
respectively. Here T_{ik} is a transformation matrix representing a reflection about the plane that is normal to the vector
The lower panel of Figure 1 shows the Pt–benzene–Pt spectral function averaged over the ensemble of junctions described in the next section using this method. Comparing the spectrum with the gasphase spectral function shown in the top panel of Figure 1, we see that screening due to the nearby Pt tips reduces the HOMO–LUMO gap by 33% on average, from 10.39 eV in the gasphase to 6.86 eV over the junction ensemble.
The screening of intramolecular Coulomb interactions by nearby conductor(s) illustrated in Figure 1 leads to an attractive interaction between a molecule and a metal surface (van der Waals interaction). By diagonalizing the molecular Hamiltonian with and without the effects of screening included in U_{nm}, it is possible to determine the van der Waals interaction at arbitrary temperature between a neutral molecule and a metallic electrode by comparing the expectation values of the Hamiltonian in these two cases:
This procedure was carried out at zero temperature for benzene oriented parallel to the surface of a planar Pt electrode at a variety of distances, and the results are shown in Figure 2. Note that an additional phenomenological shortrange repulsion proportional to r^{−12} has been included in the calculation to model the Pauli repulsion arising when the benzene πorbitals overlap the Pt surface states.
The lead–molecule coupling
When an isolated molecule is connected to electrodes and a molecular junction is formed, the energy levels of the molecule are broadened and shifted as a result of the formation of a lead–molecule bond and due to the electrostatic influence of the leads. The bonding between lead α and the molecule is described by the tunneling width matrix Γ^{α} and the electrostatics, including intramolecular screening and van der Waals effects, are described by the effective molecular Hamiltonian derived using the aforementioned πEFT. Although we use the Pt–benzene–Pt junction as an example here, the techniques we discuss are applicable to any conjugated organic molecular junction.
Bonding
The bonding between the tip of electrode α and the molecule is characterized by the tunnelingwidth matrix Γ^{α} given by Equation 4. When a highly conductive SMJ [4] is formed, the lead and molecule are in direct contact such that the overlap between the πelectron system of the molecule and all of the quasiatomic wavefunctions of the atomically sharp electrode are relevant. In this case we may express the elements of Γ^{α} as [6]
where the sum is calculated over the evanescent tunneling modes emanating from the metal tip, labeled by their angularmomentum quantum numbers, is the local density of states on the apex atom of electrode α, and is the tunneling matrix element of orbitals l [25]. The constants C_{l} can in principle be determined by matching the evanescent tip modes to the wavefunctions within the metal tip [25]; however, we set and determine the constant C by fitting to the peak of the experimental conductance histogram [4]. In the calculation of the matrix elements, we use the effective Bohr radius of a πorbital a^{*} = a_{0}/Z, where a_{0} ≈ 0.53 Å is the Bohr radius and Z = 3.22 is the effective hydrogenic charge associated with the πorbital quadrupole moment −0.65 eÅ^{2}, determined by πEFT.
For each Pt tip, we include one s, three p and five d orbitals in our calculations, which represent the evanescent tunneling modes in free space outside the apex atom of the tip. At room temperature, the Pt density of states (DOS) ρ^{α}(E) = Σ_{l}ρ_{l}^{α}(E) is sharply peaked around the Fermi energy [26] with ρ^{α}(ε_{F}) = 2.88/eV [27]. In accordance with [25], we distribute the total DOS such that the s orbital contributes 10%, the p orbitals contribute 10%, and the d orbitals contribute 80%.
We are interested in investigating transport through stable junctions where the “atop” binding configuration of benzene on Pt has the largest binding energy [2830]. In this configuration, the distance between the tip atom and the center of the benzene ring is ≈2.25 Å [4], giving a tip to orbital distance of ≈2.65 Å (the C–C bond distance is taken as 1.4 Å). The trace of Γ^{α}(ε_{F}) is shown as a function of tip position in Figure 3, where for each tip position the height was adjusted such that the distance to the closest carbon atom was 2.65 Å. From the figure, it is evident that the lead–molecule coupling strength is peaked when the tip is in the vicinity of the center of the benzene ring (whose outline is drawn schematically in black). As shown in [2], the hybridization contribution to the binding energy is
which is roughly Tr{Γ(ε_{F})}. Here μ is the chemical potential of the lead metal, is the Nparticle molecular Hilbert space, and 0_{N} is the ground state of the Nparticle manifold of the neutral molecule. The sharply peaked nature of Tr{Γ^{α}} seen in Figure 3 is thus consistent with the large binding energy of the atop configuration.
This result motivates our procedure for generating the ensemble of junctions, in which we consider the tip position in the plane parallel to the benzene ring to be a 2D Gaussian random variable with a standard deviation of 0.25 Å, chosen to correspond with the preferred bonding observed in this region. For each position, the height of each electrode (one placed above the plane and one below) is adjusted such that the closest carbon to the apex atom of each electrode is at a distance of 2.65 Å. Each lead is positioned independently of the other. This procedure ensures that the full range of possible, bonded junctions are included in the ensemble.
The eigenvalue distributions of Γ^{α} over the ensemble are shown in Figure 4. Although we include nine (orthogonal) basis orbitals for each lead, the Γ matrix only exhibits five nonzero eigenvalues, presumably because only five linear combinations can be formed that are directed toward the molecule. Although we have shown the distribution for a single lead, the number of transmission channels for two leads, where each Γ^{α} matrix has the same rank, will be the same even though the overall lead–molecule coupling strength will be larger. The average coupling per orbital with two electrodes is shown in the bottom panel of Figure 5.
Screening
The ensemble of screened interaction matrices is generated using the same procedure discussed above. Each Pt electrode is modelled as a conducting sphere with radius equal to the Pt polarization radius (1.87 Å). This is equivalent to the assumption that screening is due mainly to the apex atoms of each Pt tip. The screening surface is placed such that it lies one covalent radius away from the nearest carbon atom [14].
The average over the interaction matrix elements defines the “charging energy” of the molecule in the junction [14]. The charging energy and perorbital Tr{Γ} distributions are shown in the top and bottom panels of Figure 5, respectively, in which two electrodes are used in all calculations. As indicated by the figure, the Tr{Γ}/6 distribution is roughly four times as broad as the chargingenergy distribution. This fact justifies the use of the ensembleaverage matrix for transport calculations [2], an approximation which makes the calculation of thousands of junctions computationally tractable. The peak values of the and Tr{Γ}/6 distributions are 1.68 eV and 1.95 eV, respectively, suggesting that transport occurs in an intermediate regime in which both the particlelike and wavelike character of the charge carriers must be considered.
In addition to sampling various bonding configurations, we also consider an ensemble of junctions to sample all possible Pt surfaces. The work function of Pt ranges from 5.93 eV for the (111) surface to 5.12 eV for the (331) surface [24], and we assume that μ_{Pt} is distributed uniformly over this interval.
Using this ensemble, the conductance histogram over the ensemble of junctions can be computed, and is shown in Figure 6. The constant prefactor C appearing in the tunneling matrix elements [25] in Equation 24 was determined by fitting the peak of the calculated conductance distribution to that of the experimental conductance histogram [4]. Note that the width of the calculated conductance peak is also comparable to that of the experimental peak [4].
Results and Discussion
The transmission eigenvalue distributions for ensembles of 1.74 × 10^{5} Pt–benzene–Pt junctions calculated by using the full manybody spectrum and in the isolatedresonance approximation are shown in Figure 7a and Figure 7b, respectively. Despite the existence of five covalent bonds between the molecule and each lead (cf. Figure 4), there are only two dominant transmission channels, which arise from the twofolddegenerate HOMO resonance closest to the Pt Fermi level [2]. As proof of this point, we calculated the transmission eigenvalue distribution, over the same ensemble, using only the HOMO resonance in the isolatedresonance approximation (Equation 19). The resulting transmission eigenvalue distributions, shown in Figure 7b, are nearly identical to the full distribution shown in Figure 7a, with the exception of the small but experimentally resolvable [4] third transmission channel.
The lack of a third channel in the isolatedresonance approximation is a direct consequence of the twofold degeneracy of the HOMO resonance, which can therefore contribute at most two transmission channels. The third channel thus arises from further offresonant tunneling. In fact, we would argue that the very observation of a third channel in some Pt–benzene–Pt junctions [4] is a consequence of the very large lead–molecule coupling (~2 eV per atomic orbital) in this system. Having simulated junctions with electrodes whose DOS at the Fermi level is smaller than that of Pt, we expect junctions with Cu or Au electrodes, for example, to exhibit only two measurable transmission channels.
In order to investigate the efficacy of the isolatedresonance approximation further, we calculated the average total transmission through a Pt–benzene–Pt junction. The transmission spectra calculated using the full molecular spectrum, the isolated HOMO resonance and the isolated LUMO resonance are each shown as a function of the chemical potential of the leads μ_{Pt} in Figure 8. The spectra are averaged over 2000 bonding configurations and the blue shaded area indicates the range of possible chemical potentials for the Pt electrodes. The close correspondence between the full transmission spectrum and the isolated HOMO resonance over this range is consistent with the accuracy of the approximate method shown in Figure 7. Similarly, in the vicinity of the LUMO resonance, the isolated LUMO resonance approximation accurately characterizes the average transmission. The HOMO–LUMO asymmetry in the average transmission function arises because the HOMO resonance couples more strongly on average to the Pt tip atoms than does the LUMO resonance.
It is tempting to assume, based on the accuracy of the isolatedresonance approximation in our manybody transport theory, that an analogous “single molecular orbital” approximation would also be sufficient in a transport calculation based, e.g., on densityfunctional theory (DFT). However, this is not the case. Although the isolatedresonance approximation can also be derived within DFT, in practice, it is necessary to use an “extended molecule” to account for charge transfer between molecule and electrodes. Analyzing transport in terms of extended molecular orbitals has proven problematic. For example, the resonances of the extended molecule in [31] apparently accounted for less than 9% of the current through the junction.
Employing an extended molecule also makes it difficult, if not impossible, to interpret transport contributions in terms of the resonances of the molecule itself [31]. Since charging effects in SMJs are welldescribed in our manybody theory [6,12], there is no need to utilize an extended molecule, and therefore the resonances in our isolatedresonance approximation are true molecular resonances.
The full counting statistics of a distribution are characterized by its cumulants. By using a singleparticle theory to describe a singlechannel junction, it can be shown [32,33] that the first cumulant is related to the junction transmission function, while the second cumulant is related to the shotnoise suppression. Often this suppression is phrased in terms of the Fano factor [34]
In Figure 9 we show the distribution of F for our ensemble of junctions, where the τ_{n} have been calculated from manybody theory. Because of the fermionic character of the charge carriers, 0 ≤ F ≤ 1 , with F = 0 corresponding to completely wavelike transport, and a value of F = 1 corresponding to completely particlelike transport. From the figure, we see that F is peaked at ~0.51 implying that both the particle and wave aspects of the carriers are important, a fact which is consistent with the commensurate charging energy and bonding strength (cf. Figure 5).
In such an intermediate regime both “complementary” aspects of the charge carriers are equally important, requiring a manybody description and resulting in many subtle and interesting effects. For example, the transport in this regime displays a variety of features stemming from the interplay between Coulomb blockade and coherentinterference effects, which occur simultaneously [6,11]. Although the Fano factor reflects the nature of the transport, it is not directly related to the shotnoise power in a manybody theory. The richness of the transport in this regime, however, suggests that a full manybody calculation of a higherorder moment, such as the shot noise, may exhibit equally interesting phenomena.
Conclusion
We have developed a stateoftheart technique to model the lead–molecule coupling in highly conductive molecular junctions. The bonding between the lead and molecule was described by using an “ab initio” model in which the tunneling matrix elements between all relevant lead tip wavefunctions and the molecule were included, producing multichannel junctions naturally from a physically motivated ensemble over various contact geometries. Coulomb interactions between the molecule and the metallic leads were included by using an image multipole method within πEFT. In concert, these techniques allowed us to accurately model SMJs within our manybody theory.
The transport for an ensemble of Pt–benzene–Pt junctions, calculated by using our manybody theory, confirmed our previous finding [2] that the number of dominant transmission channels is two, with the higher channels more strongly suppressed within the more realistic treatment of leadmolecule coupling presented here. Moreover, we find that the transport through a Pt–benzene–Pt junction can be accurately described by using only the relevant (twofolddegenerate HOMO) molecular resonance. The exceptional accuracy of such an isolatedresonance approximation, however, may be limited to small molecules with large charging energies. In larger molecules, where the charging energy is smaller, further offresonant transmission channels may become more important.
In metallic point contacts the number of channels is completely determined by the valence of the metal. Despite the larger number of states available for tunneling transport in SMJs, we predict that the number of transmission channels is typically more limited than in singleatom contacts because molecules are less symmetric than atoms. Channelresolved transport measurements of SMJs therefore offer a unique probe into the symmetry of the molecular species involved.
Supporting Information
We investigate the origin of the transmission distribution widths by considering transport ensembles over Pt surfaces with fixed bonding, and over bonding configurations with a fixed Pt surface.
Supporting Information File 1: Transport distribution width decomposition  
Format: PDF  Size: 87.6 KB  Download 
References

Scheer, E.; Agraït, N.; Cuevas, J. C.; Levy Yeyati, A.; Ludoph, B.; MartínRodero, A.; Rubio Bollinger, G.; van Ruitenbeek, J. M.; Urbina, C. Nature 1998, 394, 154–157. doi:10.1038/28112
Return to citation in text: [1] 
Bergfield, J. P.; Barr, J. D.; Stafford, C. A. ACS Nano 2011, 5, 2707–2714. doi:10.1021/nn1030753
Return to citation in text: [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] 
“Molecular orbitals” are a singleparticle concept and not defined in a manybody theory. Instead we define the “HOMO resonance” as the addition spectrum peak corresponding to the N − 1 → N electronic transition of the molecule, where N is the number of electrons in the neutral molecule. Similarly, we define the “LUMO resonance” as the addition spectrum peak corresponding to the N → N + 1 transition.
Return to citation in text: [1] 
Kiguchi, M.; Tal, O.; Wohlthat, S.; Pauly, F.; Krieger, M.; Djukic, D.; Cuevas, J. C.; van Ruitenbeek, J. M. Phys. Rev. Lett. 2008, 101, 046801. doi:10.1103/PhysRevLett.101.046801
Return to citation in text: [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] 
Datta, S. Electronic Transport in Mesoscopic Systems; Cambridge University Press: Cambridge, UK, 1995; pp 117–174.
Return to citation in text: [1] 
Bergfield, J. P.; Stafford, C. A. Phys. Rev. B 2009, 79, 245125. doi:10.1103/PhysRevB.79.245125
Return to citation in text: [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] 
Toher, C.; Filippetti, A.; Sanvito, S.; Burke, K. Phys. Rev. Lett. 2005, 95, 146402. doi:10.1103/PhysRevLett.95.146402
Return to citation in text: [1] 
Koentopp, M.; Burke, K.; Evers, F. Phys. Rev. B 2006, 73, 121403. doi:10.1103/PhysRevB.73.121403
Return to citation in text: [1] 
Muralidharan, B.; Ghosh, A. W.; Datta, S. Phys. Rev. B 2006, 73, 155410. doi:10.1103/PhysRevB.73.155410
Return to citation in text: [1] 
Geskin, V.; Stadler, R.; Cornil, J. Phys. Rev. B 2009, 80, 085411. doi:10.1103/PhysRevB.80.085411
Return to citation in text: [1] [2] 
Bergfield, J. P.; Jacquod, P.; Stafford, C. A. Phys. Rev. B 2010, 82, 205405. doi:10.1103/PhysRevB.82.205405
Return to citation in text: [1] [2] 
Bergfield, J. P.; Solomon, G. C.; Stafford, C. A.; Ratner, M. A. Nano Lett. 2011, 11, 2759–2764. doi:10.1021/nl201042m
Return to citation in text: [1] [2] [3] [4] [5] 
Castleton, C. W. M.; Barford, W. J. Chem. Phys. 2002, 117, 3570–3582. doi:10.1063/1.1489994
Return to citation in text: [1] 
Barr, J. D.; Bergfield, J. P.; Stafford, C. A. Effective Field Theory of Interacting πElectrons 2011, arXiv:1108.0021v2 [condmat.meshall].
Return to citation in text: [1] [2] [3] [4] 
Djukic, D.; van Ruitenbeek, J. M. Nano Lett. 2006, 6, 789–793. doi:10.1021/nl060116e
Return to citation in text: [1] 
Solomon, G. C.; Gagliardi, A.; Pecchia, A.; Frauenheim, T.; Di Carlo, A.; Reimers, J. R.; Hush, N. S. Nano Lett. 2006, 6, 2431–2437. doi:10.1021/nl0614516
Return to citation in text: [1] 
Kovac, B.; Mohraz, M.; Heilbronner, E.; Boekelheide, V.; Hopf, H. J. Am. Chem. Soc. 1980, 102, 4314–4324. doi:10.1021/ja00533a005
Return to citation in text: [1] [2] 
Sell, J. A.; Kuppermann, A. Chem. Phys. 1978, 33, 367–378. doi:10.1016/03010104(78)870864
Return to citation in text: [1] [2] 
Kobayoshi, T. Phys. Lett. A 1978, 69, 105–108. doi:10.1016/03759601(78)903717
Return to citation in text: [1] [2] 
Schmidt, W. J. Chem. Phys. 1977, 66, 828–845. doi:10.1063/1.433961
Return to citation in text: [1] [2] 
Baltzer, P.; Karlsson, L.; Wannberg, B.; Öhrwall, G.; Holland, D. M. P.; MacDonald, M. A.; Hayes, M. A.; von Niessen, W. Chem. Phys. 1997, 224, 95–119. doi:10.1016/S03010104(97)002449
Return to citation in text: [1] 
Howell, J. O.; Goncalves, J. M.; Amatore, C.; Klasinc, L.; Wightman, R. M.; Kochi, J. K. J. Am. Chem. Soc. 1984, 106, 3968–3976. doi:10.1021/ja00326a014
Return to citation in text: [1] 
Burrow, P. D.; Michejda, J. A.; Jordan, K. D. J. Chem. Phys. 1987, 86, 9–24. doi:10.1063/1.452598
Return to citation in text: [1] 
Lide, D. R., Ed. CRC Handbook of Chemistry and Physics; CRC Press: Boca Raton, FL, 2005.
Return to citation in text: [1] [2] [3] 
Chen, C. J. Introduction to Scanning Tunneling Microscopy, 2nd ed.; Oxford University Press: New York, 1993.
Return to citation in text: [1] [2] [3] [4] 
Kleber, R. Z. Phys. A: Hadrons Nucl. 1973, 264, 301–308. doi:10.1007/BF01398856
Return to citation in text: [1] 
Kittel, C. Introduction to solid state physics, 5th ed.; John Wiley and Sons, Inc.: New York, 1976.
Return to citation in text: [1] 
Cruz, M. T. d. M.; Carneiro, J. W. d. M.; Aranda, D. A. G.; Bühl, M. J. Phys. Chem. C 2007, 111, 11068–11076. doi:10.1021/jp072572c
Return to citation in text: [1] 
Morin, C.; Simon, D.; Sautet, P. J. Phys. Chem. B 2003, 107, 2995–3002. doi:10.1021/jp026950j
Return to citation in text: [1] 
Saeys, M.; Reyniers, M.F.; Marin, G. B.; Neurock, M. J. Phys. Chem. B 2002, 106, 7489–7498. doi:10.1021/jp0201231
Return to citation in text: [1] 
Heurich, J.; Cuevas, J. C.; Wenzel, W.; Schön, G. Phys. Rev. Lett. 2002, 88, 256803. doi:10.1103/PhysRevLett.88.256803
Return to citation in text: [1] [2] 
Levitov, L. S.; Lee, H.; Lesovik, G. B. J. Math. Phys. 1996, 37, 4845–4866. doi:10.1063/1.531672
Return to citation in text: [1] 
Levitov, L. S.; Lesovik, G. B. JETP Lett. 1993, 58, 230.
http://adsabs.harvard.edu/abs/1993JETPL..58..230L
Return to citation in text: [1] 
Büttiker, M. Phys. Rev. Lett. 1990, 65, 2901–2904. doi:10.1103/PhysRevLett.65.2901
Return to citation in text: [1]
6.  Bergfield, J. P.; Stafford, C. A. Phys. Rev. B 2009, 79, 245125. doi:10.1103/PhysRevB.79.245125 
12.  Bergfield, J. P.; Solomon, G. C.; Stafford, C. A.; Ratner, M. A. Nano Lett. 2011, 11, 2759–2764. doi:10.1021/nl201042m 
14.  Barr, J. D.; Bergfield, J. P.; Stafford, C. A. Effective Field Theory of Interacting πElectrons 2011, arXiv:1108.0021v2 [condmat.meshall]. 
17.  Kovac, B.; Mohraz, M.; Heilbronner, E.; Boekelheide, V.; Hopf, H. J. Am. Chem. Soc. 1980, 102, 4314–4324. doi:10.1021/ja00533a005 
18.  Sell, J. A.; Kuppermann, A. Chem. Phys. 1978, 33, 367–378. doi:10.1016/03010104(78)870864 
19.  Kobayoshi, T. Phys. Lett. A 1978, 69, 105–108. doi:10.1016/03759601(78)903717 
20.  Schmidt, W. J. Chem. Phys. 1977, 66, 828–845. doi:10.1063/1.433961 
21.  Baltzer, P.; Karlsson, L.; Wannberg, B.; Öhrwall, G.; Holland, D. M. P.; MacDonald, M. A.; Hayes, M. A.; von Niessen, W. Chem. Phys. 1997, 224, 95–119. doi:10.1016/S03010104(97)002449 
25.  Chen, C. J. Introduction to Scanning Tunneling Microscopy, 2nd ed.; Oxford University Press: New York, 1993. 
4.  Kiguchi, M.; Tal, O.; Wohlthat, S.; Pauly, F.; Krieger, M.; Djukic, D.; Cuevas, J. C.; van Ruitenbeek, J. M. Phys. Rev. Lett. 2008, 101, 046801. doi:10.1103/PhysRevLett.101.046801 
6.  Bergfield, J. P.; Stafford, C. A. Phys. Rev. B 2009, 79, 245125. doi:10.1103/PhysRevB.79.245125 
25.  Chen, C. J. Introduction to Scanning Tunneling Microscopy, 2nd ed.; Oxford University Press: New York, 1993. 
24.  Lide, D. R., Ed. CRC Handbook of Chemistry and Physics; CRC Press: Boca Raton, FL, 2005. 
4.  Kiguchi, M.; Tal, O.; Wohlthat, S.; Pauly, F.; Krieger, M.; Djukic, D.; Cuevas, J. C.; van Ruitenbeek, J. M. Phys. Rev. Lett. 2008, 101, 046801. doi:10.1103/PhysRevLett.101.046801 
17.  Kovac, B.; Mohraz, M.; Heilbronner, E.; Boekelheide, V.; Hopf, H. J. Am. Chem. Soc. 1980, 102, 4314–4324. doi:10.1021/ja00533a005 
18.  Sell, J. A.; Kuppermann, A. Chem. Phys. 1978, 33, 367–378. doi:10.1016/03010104(78)870864 
19.  Kobayoshi, T. Phys. Lett. A 1978, 69, 105–108. doi:10.1016/03759601(78)903717 
20.  Schmidt, W. J. Chem. Phys. 1977, 66, 828–845. doi:10.1063/1.433961 
22.  Howell, J. O.; Goncalves, J. M.; Amatore, C.; Klasinc, L.; Wightman, R. M.; Kochi, J. K. J. Am. Chem. Soc. 1984, 106, 3968–3976. doi:10.1021/ja00326a014 
23.  Burrow, P. D.; Michejda, J. A.; Jordan, K. D. J. Chem. Phys. 1987, 86, 9–24. doi:10.1063/1.452598 
26.  Kleber, R. Z. Phys. A: Hadrons Nucl. 1973, 264, 301–308. doi:10.1007/BF01398856 
27.  Kittel, C. Introduction to solid state physics, 5th ed.; John Wiley and Sons, Inc.: New York, 1976. 
25.  Chen, C. J. Introduction to Scanning Tunneling Microscopy, 2nd ed.; Oxford University Press: New York, 1993. 
24.  Lide, D. R., Ed. CRC Handbook of Chemistry and Physics; CRC Press: Boca Raton, FL, 2005. 
25.  Chen, C. J. Introduction to Scanning Tunneling Microscopy, 2nd ed.; Oxford University Press: New York, 1993. 
14.  Barr, J. D.; Bergfield, J. P.; Stafford, C. A. Effective Field Theory of Interacting πElectrons 2011, arXiv:1108.0021v2 [condmat.meshall]. 
2.  Bergfield, J. P.; Barr, J. D.; Stafford, C. A. ACS Nano 2011, 5, 2707–2714. doi:10.1021/nn1030753 
2.  Bergfield, J. P.; Barr, J. D.; Stafford, C. A. ACS Nano 2011, 5, 2707–2714. doi:10.1021/nn1030753 
14.  Barr, J. D.; Bergfield, J. P.; Stafford, C. A. Effective Field Theory of Interacting πElectrons 2011, arXiv:1108.0021v2 [condmat.meshall]. 
28.  Cruz, M. T. d. M.; Carneiro, J. W. d. M.; Aranda, D. A. G.; Bühl, M. J. Phys. Chem. C 2007, 111, 11068–11076. doi:10.1021/jp072572c 
29.  Morin, C.; Simon, D.; Sautet, P. J. Phys. Chem. B 2003, 107, 2995–3002. doi:10.1021/jp026950j 
30.  Saeys, M.; Reyniers, M.F.; Marin, G. B.; Neurock, M. J. Phys. Chem. B 2002, 106, 7489–7498. doi:10.1021/jp0201231 
4.  Kiguchi, M.; Tal, O.; Wohlthat, S.; Pauly, F.; Krieger, M.; Djukic, D.; Cuevas, J. C.; van Ruitenbeek, J. M. Phys. Rev. Lett. 2008, 101, 046801. doi:10.1103/PhysRevLett.101.046801 
4.  Kiguchi, M.; Tal, O.; Wohlthat, S.; Pauly, F.; Krieger, M.; Djukic, D.; Cuevas, J. C.; van Ruitenbeek, J. M. Phys. Rev. Lett. 2008, 101, 046801. doi:10.1103/PhysRevLett.101.046801 
4.  Kiguchi, M.; Tal, O.; Wohlthat, S.; Pauly, F.; Krieger, M.; Djukic, D.; Cuevas, J. C.; van Ruitenbeek, J. M. Phys. Rev. Lett. 2008, 101, 046801. doi:10.1103/PhysRevLett.101.046801 
4.  Kiguchi, M.; Tal, O.; Wohlthat, S.; Pauly, F.; Krieger, M.; Djukic, D.; Cuevas, J. C.; van Ruitenbeek, J. M. Phys. Rev. Lett. 2008, 101, 046801. doi:10.1103/PhysRevLett.101.046801 
1.  Scheer, E.; Agraït, N.; Cuevas, J. C.; Levy Yeyati, A.; Ludoph, B.; MartínRodero, A.; Rubio Bollinger, G.; van Ruitenbeek, J. M.; Urbina, C. Nature 1998, 394, 154–157. doi:10.1038/28112 
2.  Bergfield, J. P.; Barr, J. D.; Stafford, C. A. ACS Nano 2011, 5, 2707–2714. doi:10.1021/nn1030753 
14.  Barr, J. D.; Bergfield, J. P.; Stafford, C. A. Effective Field Theory of Interacting πElectrons 2011, arXiv:1108.0021v2 [condmat.meshall]. 
31.  Heurich, J.; Cuevas, J. C.; Wenzel, W.; Schön, G. Phys. Rev. Lett. 2002, 88, 256803. doi:10.1103/PhysRevLett.88.256803 
4.  Kiguchi, M.; Tal, O.; Wohlthat, S.; Pauly, F.; Krieger, M.; Djukic, D.; Cuevas, J. C.; van Ruitenbeek, J. M. Phys. Rev. Lett. 2008, 101, 046801. doi:10.1103/PhysRevLett.101.046801 
4.  Kiguchi, M.; Tal, O.; Wohlthat, S.; Pauly, F.; Krieger, M.; Djukic, D.; Cuevas, J. C.; van Ruitenbeek, J. M. Phys. Rev. Lett. 2008, 101, 046801. doi:10.1103/PhysRevLett.101.046801 
3.  “Molecular orbitals” are a singleparticle concept and not defined in a manybody theory. Instead we define the “HOMO resonance” as the addition spectrum peak corresponding to the N − 1 → N electronic transition of the molecule, where N is the number of electrons in the neutral molecule. Similarly, we define the “LUMO resonance” as the addition spectrum peak corresponding to the N → N + 1 transition. 
13.  Castleton, C. W. M.; Barford, W. J. Chem. Phys. 2002, 117, 3570–3582. doi:10.1063/1.1489994 
4.  Kiguchi, M.; Tal, O.; Wohlthat, S.; Pauly, F.; Krieger, M.; Djukic, D.; Cuevas, J. C.; van Ruitenbeek, J. M. Phys. Rev. Lett. 2008, 101, 046801. doi:10.1103/PhysRevLett.101.046801 
2.  Bergfield, J. P.; Barr, J. D.; Stafford, C. A. ACS Nano 2011, 5, 2707–2714. doi:10.1021/nn1030753 
4.  Kiguchi, M.; Tal, O.; Wohlthat, S.; Pauly, F.; Krieger, M.; Djukic, D.; Cuevas, J. C.; van Ruitenbeek, J. M. Phys. Rev. Lett. 2008, 101, 046801. doi:10.1103/PhysRevLett.101.046801 
24.  Lide, D. R., Ed. CRC Handbook of Chemistry and Physics; CRC Press: Boca Raton, FL, 2005. 
7.  Toher, C.; Filippetti, A.; Sanvito, S.; Burke, K. Phys. Rev. Lett. 2005, 95, 146402. doi:10.1103/PhysRevLett.95.146402 
8.  Koentopp, M.; Burke, K.; Evers, F. Phys. Rev. B 2006, 73, 121403. doi:10.1103/PhysRevB.73.121403 
9.  Muralidharan, B.; Ghosh, A. W.; Datta, S. Phys. Rev. B 2006, 73, 155410. doi:10.1103/PhysRevB.73.155410 
10.  Geskin, V.; Stadler, R.; Cornil, J. Phys. Rev. B 2009, 80, 085411. doi:10.1103/PhysRevB.80.085411 
2.  Bergfield, J. P.; Barr, J. D.; Stafford, C. A. ACS Nano 2011, 5, 2707–2714. doi:10.1021/nn1030753 
4.  Kiguchi, M.; Tal, O.; Wohlthat, S.; Pauly, F.; Krieger, M.; Djukic, D.; Cuevas, J. C.; van Ruitenbeek, J. M. Phys. Rev. Lett. 2008, 101, 046801. doi:10.1103/PhysRevLett.101.046801 
2.  Bergfield, J. P.; Barr, J. D.; Stafford, C. A. ACS Nano 2011, 5, 2707–2714. doi:10.1021/nn1030753 
6.  Bergfield, J. P.; Stafford, C. A. Phys. Rev. B 2009, 79, 245125. doi:10.1103/PhysRevB.79.245125 
11.  Bergfield, J. P.; Jacquod, P.; Stafford, C. A. Phys. Rev. B 2010, 82, 205405. doi:10.1103/PhysRevB.82.205405 
12.  Bergfield, J. P.; Solomon, G. C.; Stafford, C. A.; Ratner, M. A. Nano Lett. 2011, 11, 2759–2764. doi:10.1021/nl201042m 
4.  Kiguchi, M.; Tal, O.; Wohlthat, S.; Pauly, F.; Krieger, M.; Djukic, D.; Cuevas, J. C.; van Ruitenbeek, J. M. Phys. Rev. Lett. 2008, 101, 046801. doi:10.1103/PhysRevLett.101.046801 
6.  Bergfield, J. P.; Stafford, C. A. Phys. Rev. B 2009, 79, 245125. doi:10.1103/PhysRevB.79.245125 
2.  Bergfield, J. P.; Barr, J. D.; Stafford, C. A. ACS Nano 2011, 5, 2707–2714. doi:10.1021/nn1030753 
5.  Datta, S. Electronic Transport in Mesoscopic Systems; Cambridge University Press: Cambridge, UK, 1995; pp 117–174. 
6.  Bergfield, J. P.; Stafford, C. A. Phys. Rev. B 2009, 79, 245125. doi:10.1103/PhysRevB.79.245125 
4.  Kiguchi, M.; Tal, O.; Wohlthat, S.; Pauly, F.; Krieger, M.; Djukic, D.; Cuevas, J. C.; van Ruitenbeek, J. M. Phys. Rev. Lett. 2008, 101, 046801. doi:10.1103/PhysRevLett.101.046801 
2.  Bergfield, J. P.; Barr, J. D.; Stafford, C. A. ACS Nano 2011, 5, 2707–2714. doi:10.1021/nn1030753 
15.  Djukic, D.; van Ruitenbeek, J. M. Nano Lett. 2006, 6, 789–793. doi:10.1021/nl060116e 
16.  Solomon, G. C.; Gagliardi, A.; Pecchia, A.; Frauenheim, T.; Di Carlo, A.; Reimers, J. R.; Hush, N. S. Nano Lett. 2006, 6, 2431–2437. doi:10.1021/nl0614516 
4.  Kiguchi, M.; Tal, O.; Wohlthat, S.; Pauly, F.; Krieger, M.; Djukic, D.; Cuevas, J. C.; van Ruitenbeek, J. M. Phys. Rev. Lett. 2008, 101, 046801. doi:10.1103/PhysRevLett.101.046801 
32.  Levitov, L. S.; Lee, H.; Lesovik, G. B. J. Math. Phys. 1996, 37, 4845–4866. doi:10.1063/1.531672 
33. 
Levitov, L. S.; Lesovik, G. B. JETP Lett. 1993, 58, 230.
http://adsabs.harvard.edu/abs/1993JETPL..58..230L 
34.  Büttiker, M. Phys. Rev. Lett. 1990, 65, 2901–2904. doi:10.1103/PhysRevLett.65.2901 
31.  Heurich, J.; Cuevas, J. C.; Wenzel, W.; Schön, G. Phys. Rev. Lett. 2002, 88, 256803. doi:10.1103/PhysRevLett.88.256803 
6.  Bergfield, J. P.; Stafford, C. A. Phys. Rev. B 2009, 79, 245125. doi:10.1103/PhysRevB.79.245125 
12.  Bergfield, J. P.; Solomon, G. C.; Stafford, C. A.; Ratner, M. A. Nano Lett. 2011, 11, 2759–2764. doi:10.1021/nl201042m 
6.  Bergfield, J. P.; Stafford, C. A. Phys. Rev. B 2009, 79, 245125. doi:10.1103/PhysRevB.79.245125 
6.  Bergfield, J. P.; Stafford, C. A. Phys. Rev. B 2009, 79, 245125. doi:10.1103/PhysRevB.79.245125 
6.  Bergfield, J. P.; Stafford, C. A. Phys. Rev. B 2009, 79, 245125. doi:10.1103/PhysRevB.79.245125 
12.  Bergfield, J. P.; Solomon, G. C.; Stafford, C. A.; Ratner, M. A. Nano Lett. 2011, 11, 2759–2764. doi:10.1021/nl201042m 
6.  Bergfield, J. P.; Stafford, C. A. Phys. Rev. B 2009, 79, 245125. doi:10.1103/PhysRevB.79.245125 
10.  Geskin, V.; Stadler, R.; Cornil, J. Phys. Rev. B 2009, 80, 085411. doi:10.1103/PhysRevB.80.085411 
6.  Bergfield, J. P.; Stafford, C. A. Phys. Rev. B 2009, 79, 245125. doi:10.1103/PhysRevB.79.245125 
12.  Bergfield, J. P.; Solomon, G. C.; Stafford, C. A.; Ratner, M. A. Nano Lett. 2011, 11, 2759–2764. doi:10.1021/nl201042m 
2.  Bergfield, J. P.; Barr, J. D.; Stafford, C. A. ACS Nano 2011, 5, 2707–2714. doi:10.1021/nn1030753 
6.  Bergfield, J. P.; Stafford, C. A. Phys. Rev. B 2009, 79, 245125. doi:10.1103/PhysRevB.79.245125 
11.  Bergfield, J. P.; Jacquod, P.; Stafford, C. A. Phys. Rev. B 2010, 82, 205405. doi:10.1103/PhysRevB.82.205405 
6.  Bergfield, J. P.; Stafford, C. A. Phys. Rev. B 2009, 79, 245125. doi:10.1103/PhysRevB.79.245125 
2.  Bergfield, J. P.; Barr, J. D.; Stafford, C. A. ACS Nano 2011, 5, 2707–2714. doi:10.1021/nn1030753 
© 2012 Bergfield et al; licensee BeilsteinInstitut.
This is an Open Access article under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The license is subject to the Beilstein Journal of Nanotechnology terms and conditions: (http://www.beilsteinjournals.org/bjnano)