Counterion-Dependent Charge Transport in Ultrathin-Film Molecular Junctions

Abstract

Counterions are ubiquitous in charged molecular films, yet their influence on charge transport in solid-state molecular junctions (MJs) remains insufficiently understood. Here, we investigate how embedded counterions regulate conductance in MJs based on ultrathin molecular films (2–9 nm) of 4,4′-Diamino-2,2′-stilbenedisulfonic acid (DAS). These DAS films were electrochemically grafted onto Au electrodes via in situ generated diazonium precursors. During film growth, cations from the supporting electrolyte are incorporated into the molecular network through electrostatic interactions with the sulfonate groups. The dense structure of the electro-grafted films enables direct deposition of Ti/Au top contacts, yielding robust large-area junctions with well-defined thicknesses. By varying the incorporated counterions (H⁺, Li⁺, Na⁺, or K⁺), we show that junction conductance is strongly counterion dependent, whereas the transport attenuation factor remains nearly unchanged. This behavior indicates that counterions primarily affect the prefactor J ₀, consistent with a counterion-dependent change in interfacial coupling. Na⁺- and K⁺-containing films favor a large population of strongly coupled sulfonate-metal pathways, leading to markedly higher conductance than H⁺- and Li⁺-containing MJs. These results identify embedded ionic species as active chemical regulators of molecule-electrode coupling and provide a practical strategy for tuning charge transport in molecular electronic devices.

Introduction

The molecular junction (MJ), in which a single molecule or a molecular ensemble bridges two conductive electrodes, is a fundamental building block in molecular electronics. Since its emergence in the latter half of the 20th century, it has attracted sustained interest as an ideal platform for probing charge transport at the molecular scale. − Charge transport in such two-terminal devices has been extensively investigated and reviewed, − and is strongly governed by the electrode separation distance (d). At short distances, coherent tunneling is usually the dominant transport process. In this regime, electrons traverse an energy barrier quantum mechanically, and the tunneling current density (J) decreases exponentially with distance according to Simmons’ model, commonly expressed in simplified form as J = J ₀ e^–βd , where J ₀ is the extrapolated current density at d = 0, and β is the attenuation factor.

β values are usually between 2 and 3 nm^–1 for conjugated molecules. ,− However, substantially smaller β values, below 1 nm^–1, have also been reported, particularly in long-range charge transport systems. − Such behavior is generally associated with near-resonant tunneling, in which a molecular electronic state lies close to the electrode Fermi level, thereby lowering the effective tunneling barrier¹². Viologen-based MJs, for instance, exhibit long-range charge transport with an unusually small attenuation factor (β ≈ 0.25 nm^–1) over distances of 3–14 nm, which has been attributed to the low-lying viologen LUMO level being positioned between the Fermi levels of the two contacts. Similarly, cobalt-terpyridine-based MJs show a β value as low as 0.17 nm^–1 because the HOMO lies between the Fermi levels of Au and Ti contacts¹⁷.

In ion-containing MJ systems, ions and counterions can modulate charge transport through ion migration, ion–molecule interactions, − and ion-induced changes in interfacial coupling. , In bipyridinium-based single-molecule junctions, conductance is dependent on the chemical identity of the counterions. More recently, enhanced conductance was reported in large-area MJs formed by intercalating transition-metal ions into self-assembled bipyridine monolayers. Despite these advances, the mechanisms by which embedded counterions modulate charge transport in solid-state large-area MJs remain unresolved.

In this work, we systematically investigated the influence of incorporated counterions on charge transport in solid-state large-area MJs. The molecular building block employed is 4,4′-diamino-2,2′-stilbenedisulfonic acid (DAS), whose chemical structure is shown in Figure a.

1.

Electro-grafting of DAS and structure of the cross-bar junction. (a) Chemical structure of 4,4′-Diamino-2,2′-stilbenedisulfonic acid (DAS). (b) Scheme of the electro-reduction of DAS-derived diazonium salts, leading to the formation of direct C-surface/C–C bonding and/or an azo bridge. (c) Schematic illustration of a plausible DAS layer composition on the Au bottom electrode, as deduced from XPS analysis. X = H, Li, Na, or K. (d) Optical micrograph of a fabricated cross-bar molecular junction. The vertical strip is the DAS-modified Au bottom electrode, and the horizontal strip is the Ti/Au top electrode. The overlap between the two electrodes defines a 20 μm × 20 μm junction area.

Results

Ultrathin DAS films (2–9 nm thick) were electrochemically grafted onto gold strip electrodes via the in situ generation and reduction of diazonium precursors (Figure b). − This process yields compact, pinhole-free molecular films, as supported by electrochemical passivation experiments (Figure S1a,b). During film growth, cations from the supporting electrolyte are simultaneously incorporated into the DAS layers to compensate the sulfonate groups and maintain electroneutrality (Figure c).

Film thicknesses were determined by atomic force microscopy (AFM) using an automated Python-based data-processing workflow described in the Supporting Information. The compactness of the electro-grafted DAS layers enables direct deposition of top metal contacts without substantial metal penetration, , thereby yielding the cross-bar device illustrated in Figure d. The detailed fabrication protocol has been reported previously. The top electrode consists of 5 nm Ti followed by 50 nm of Au. The Ti interlayer promotes interfacial coupling and suppresses Au penetration. The resulting molecular junctions are denoted Au₅₀/DAS_X_t/Ti₅/Au₅₀, where subscripts indicate the layer thicknesses in nanometers, and X is the countercation incorporated during film growth.

To introduce different counterions into DAS films, HClO₄, LiClO₄, NaClO₄, and KClO₄ were used as supporting electrolytes during electro-grafting. For clarity, DAS layers grown in XClO₄ electrolytes (X = H, Li, Na, or K) are denoted DAS_X. After grafting, the chemical composition of the resulting films was characterized by X-ray photoelectron spectroscopy (XPS).

The first important result from the XPS analysis is the total N/S ratio (Table ), which provides insight into the chemical structure of the grafted DAS layers. The DAS precursor contains two amino groups and two sulfonic acid groups (Figure a), corresponding to an initial N/S ratio of 1. During diazotization and electrochemical reduction, amino groups are converted into diazonium groups, which are expected to release N₂ upon reduction (Figure b, upper pathway). In an ideal fully reduced grafted layer in which both amine groups of each DAS unit are eliminated, no nitrogen signal would remain in the film. If only one amino group per DAS unit is diazotized and participates in surface grafting or oligomer growth while the other amino group remains intact, an N/S ratio close to 0.5 would be expected. The experimentally measured N/S ratios range from 0.76 to 1.15 (Table ), systematically exceeding this limiting value. This indicates that, in addition to residual amino species, azo bridges are formed during the film growth (Figure b, lower pathway). Such azo coupling reactions are known to compete with direct radical grafting during diazonium reduction and can introduce additional nitrogen atoms into the molecular layer. , Consistent with this interpretation, DAS_H, DAS_Li, DAS_Na, and DAS_K films exhibit similar N_1s spectra (Figures a,b, andS4), which can be deconvoluted into a component near 400 eV which is classically attributed to azo (−NN−) nitrogen and a higher-binding-energy component near 402 eV. However, in the case of our DAS-based layers, the quantitative interpretation of the signal at 400 eV is complicated by the specific molecular structure of the grafted species. As the DAS molecule possesses two amine functions (−NH₂), and the N_1s binding energy of neutral amine nitrogen (C–NH₂) also falls in the vicinity of 400 eV, The literature − itself acknowledges this ambiguity  the 400 eV peak has been variously assigned to azo bridges, “reduced nitrogen,” or amino groups. So the entire 400 eV component cannot be attributed solely to azo bridges; doing so would overestimate the azo fraction. In the present case the number of remaining nitrogens in the films after electrochemical grafting is quite high indicating a large fraction of azo-link or remaining amino groups when compared to literature − and a combined approach using both Raman and FTIR would therefore provide a most complete picture.

1. Elemental Composition and Cation-to-Sulfur Ratios of DAS_X Films (X = H, Li, Na, or K) Determined by XPS.

Sample	N/S	N+/S	X+/S	Total cation/S	BE of N+ (eV)
DAS_H	1.15	0.74	0.26	-	401.66
DAS_Li	0.84	0.36	0.64	-	402.06
DAS_Na	0.98	0.39	0.62	1.01	402.01
DAS_K	0.76	0.32	0.68	1.00	401.97

^aEstimated from charge neutrality, assuming total cation/S = 1.

2.

XPS characterization of ultrathin DAS films on Au substrates: N_1s high-resolution spectra of films grown in (a) HClO₄ and (b) LiClO₄; (c) high-resolution Na signal of film grown in NaClO₄; and (d) high-resolution K signal of film grown in KClO₄.

The higher-binding-energy component is assigned to positively charged nitrogen sites. In DAS_Li, DAS_Na, and DAS_K films, this contribution is most plausibly dominated by protonated residual amine sites. In the DAS_H film grown in 0.1 M HClO₄, protonation of azo linkages is also possible and can contribute to the N⁺ signal under strongly acidic conditions. The DAS_H films exhibit a distinct N⁺ component at 401.66 eV (Figure a), slightly lower than the corresponding components in the alkali-metal-containing films (Table ), which is consistent with the contribution of protonated azo species, where the sp2 hybridization lowers the binding energy of the N_1s core level.

The second important result from XPS is the direct evidence for counterion incorporation into DAS films. As shown in Figure c,d, clear Na and K signals are observed for DAS_Na and DAS_K films, respectively, with Na/S and K/S ratios of 0.62 and 0.68 (Table ). These values confirm efficient incorporation of alkali-metal counterions from the supporting electrolyte. The charge balance derived from XPS is summarized in Table . For DAS_Na and DAS_K, the total cation/S ratios, calculated as the sum of the positively charged nitrogen sites and the alkali-metal ions divided by the sulfur content, are approximately 1.01 and 1.00, respectively, in agreement with the electroneutrality of the sulfonate-containing layer. This observation motivates extending the assumption to the DAS_H and DAS_Li samples, in which Li cannot be quantified reliably by XPS and H⁺ is not directly detectable by XPS; therefore, the Li/S ratio was estimated as 0.64 from charge neutrality, consistent with the Na/S and K/S values. For DAS_H, the N⁺/S ratio is much larger (0.74) than in the alkali-metal-containing films (0.32–0.39), indicating a higher concentration of protonated nitrogen sites in the acid-grown film due to protonated azo groups. The remaining charge required to compensate the sulfonate groups corresponds to an estimated H⁺/S ratio of 0.26.

To examine the role of incorporated ions in charge transport, molecular junctions were fabricated by depositing the metallic top electrode onto the DAS-modified Au strip electrodes. Current density–voltage (J–V) characteristics were measured with the top electrode grounded. Thus, the applied voltage corresponds to the potential of the bottom Au electrode relative to the top Ti/Au electrode, and positive current denotes current flow from the bottom electrode through the molecular layer to the top electrode. Figure a presents the J–V characteristic of 11 Au/DAS_H_3.3/Ti/Au devices, with the average curve shown in bold black. The narrow current dispersion indicates homogeneous DAS_H film formation and a reproducible junction fabrication. Figure b shows representative J–V curves for Au/DAS_H_t/Ti/Au junctions with DAS_H thicknesses ranging from 2.2 to 5.6 nm (see Figure S5 for all data set). All devices exhibit slightly asymmetric J–V characteristics, with higher current at positive bias. This asymmetry mainly arises from the asymmetry of the two contacts. In the Supporting Information, all data set for the others cations have been presented (Figures S6– S8).

3.

Current density–voltage (J–V) characteristics of DAS-based molecular junctions. (a) J–V curves of 11 Au/DAS_H_3.3/Ti/Au devices, with the average shown in bold. (b) J–V curves of Au/DAS_H/Ti/Au junctions with different DAS_H film thicknesses. (c) Comparison of J–V characteristics for Au/DAS_H_5.6/Ti/Au, Au/DAS_Li_6.0/Ti/Au, Au/DAS_Na_6.2/Ti/Au, and Au/DAS_K_6.3/Ti/Au junctions with similar DAS layer thicknesses. (d) Semilogarithmic plot of ln J at 1 V as a function of DAS thickness (nm) for different counterions.

Figure c compares representatives J–V curves for junctions incorporating different counterions while maintaining similar film thicknesses (approximately 6 nm). A pronounced dependence of junction conductivity on the counterion identity is observed. The current density decreases in the order DAS_Na > DAS_K > DAS_H ≈ DAS_Li, with current ratios of approximately 70:22:2.6:1 at 1 V when normalized to the DAS_Li device. Minor variations in film thickness are unavoidable because diazonium-mediated electro-grafting does not allow exact thickness control. For example, the thickness difference between DAS_H and DAS_K in Figure c is 0.7 nm (Table S1), corresponding roughly to the length of a single DAS unit. However, such a small difference in thickness cannot account for the full observed conductance variation, which exceeds 1 order of magnitude. Furthermore, extensive AFM characterization (Figure S3 and Table S2) was performed to demonstrate that the structure of the electrografted molecular layers is not affected by the incorporated cations. Even though differences in molecular packing density or layer structure arising from the incorporation of different counter-cations could, in principle, contribute to the observed conductance variations, the combined AFM morphology, quantitative roughness analysis, and electrical trends argue against substantial differences in molecular density or film morphology being the dominant origin of the observed counterion-dependent conductance.

To further confirm the counterion effect, junctions with different DAS thicknesses and counterions were fabricated and analyzed. Figure d shows the semilogarithmic plot of ln J at 1 V as a function of film thickness for different DAS_X junctions. DAS_H- and DAS_Li-based junctions exhibit similarly low current densities, whereas DAS_K and DAS_Na devices are much more conductive. Linear fits of ln J versus thickness yield attenuation factors β values close to 1 nm^–1 for all counterions, indicating a common attenuation regime and suggesting that the dominant charge-transport mechanism is not affected by the counterion identity. It is important to note that the conductance difference in Figure c is much larger than the one expected from the thickness variation. From the Simmons relation, for β ≈ 1 nm^–1, a thickness difference of 0.7 nm produces a current-density variation of only approximately a factor of 2, which is clearly lower than the observed ratio.

Discussion

Electroactivity characterization of the DAS layer (Figure S1c) and DFT calculation (Figure S1d) together indicate that the HOMO level of DAS oligomer is approximately −5.6 eV, similar to that of oligo-bisthienylbenzene (BTB) layers. A transition in the attenuation plot has been reported for BTB-based MJs at 5 nm. In contrast, no such transition is observed in DAS-based MJs over the thickness range investigated here. This behavior is tentatively attributed to the ionic screening at DAS/metal interfaces, which can alter the internal electric field and shift the potential profile and molecular orbital energies (Figure a,b). , Therefore, the relatively small β value and the absence of a clear transition between 2 and 9 nm can be rationalized by a reduced effective tunneling barrier, arising from the proximity of DAS HOMO to the Au Fermi level, together with the ionic screening. This combination favors near-resonant tunneling over comparatively long distances. ,

4.

Energetic landscape across the MJ and schematic illustration of counterion-dependent interfacial coupling. Energy level alignment in the absence of mobile ions in BTB (a) and in the presence of embedded/mobile ions in DAS (b), where + denotes the mobile cations and – denotes the immobilized SO₃ ^– groups. Note: 1 V is applied to the Au electrode. (c) Illustration of a weak interfacial coupling between the DAS layer and metal contact, proposed to dominate in DAS_H- and DAS-Li-based MJs. (d) Stronger interfacial coupling with metal-sulfonate interactions, proposed to be more prevalent in DAS_Na- and DAS-K-based MJs.

The main counterion-dependent parameter is therefore J ₀. DAS_Na-based junctions exhibit ln J ₀ ≈ 1.9, whereas DAS_H-based junctions show ln J ₀ of −2.2, corresponding to a roughly 55-fold difference in J ₀. Differences in J ₀ at constant β have been widely reported. Evers and co-workers demonstrated that single-molecule Au/alkanedithiol/Au junctions can display multiple conductance states with identical attenuation factors but different J ₀ values, arising from different molecular binding geometries. Venkataraman and co-workers later showed that stronger Au–C covalent coupling can increase single-molecular conductance by 2 orders of magnitude compared with Au-NH₂ interaction, while preserving exponential length dependence for longer oligomers; recent theoretical calculations also support this conclusion. In large-area molecular junctions, electric measurements reflect the ensemble response of numerous parallel conduction pathways, making single-molecule arrangement effects less pronounced. Nevertheless, the same physical principle applies: the conductance prefactor is sensitive to the electrode-molecule coupling. In addition, J ₀ in large-area MJs can also be related to the density of active molecular wires within the junction.

To rationalize the counterion-dependent J ₀ observed here, counterion size and ion-pair strength must be considered. All counterions used in this study are monovalent and exhibit broadly similar spherical structures, but they differ in ionic radius, ion-pairing strength, and mobility within confined molecular environments. The Anderson-Stuart model for ionic motion in solids predicts that the activation barrier for ion displacement depends strongly on ionic radius and local free volume. This model suggests that Na⁺ can display particularly favorable mobility/lability, whereas Li⁺ is more strongly bound and K⁺ requires a larger local free volume. This qualitative trend is consistent with the conductance order observed in this study, although the measured electronic current is not attributed to direct ionic conduction.

Instead, the counterion dependence of J ₀ is best rationalized by a counterion-controlled interfacial coupling effect. In all DAS_X films, sulfonate groups are the only anionic sites and are distributed throughout the grafted molecular network (Figure c). At the molecule/metal interface, the protonation state of acid anchoring groups can strongly affect contact resistance and molecular conductance. We propose that H⁺ and Li⁺ produce more strongly associated ion pairs or protonated sites, which lowers the average availability of ionic or polar interfacial sites for efficient coupling to the top contact. By contrast, Na⁺ and K⁺ form more labile ion-pair environments, allowing counterion displacement and increasing the probability of electronically efficient interfacial configurations (Figure c,d). The higher J ₀ values observed for DAS_Na and DAS_K junctions are therefore assigned to a larger population of strongly coupled interfacial pathways. This interpretation is consistent with the similar β values measured for all DAS_X junctions: counterions do not substantially modify the molecular attenuation pathway, but they regulate the number and strength of electronically efficient pathways.

Recent studies further support the view that ions can regulate molecular conductance by modulating interfacial coupling. Tong and co-workers demonstrated a local-cation-controlled single-molecule switch in which metal cations in the electric double layer modulate Au–carboxylate contacts, producing an On/Off ratio exceeding 1000. In that system, Na⁺, K⁺, Mg²⁺, and Ca²⁺ coordinated with deprotonated carboxylate groups, suppressing the formation of Au–carboxylate contacts and switching the junction to a low-conductance state. Although the present DAS junctions are solid-state ensemble devices, the same general principle is relevant: the local ionic environment can modulate the accessibility of interfacial binding motifs that promote efficient electronic coupling. The conductance modulation observed here is smaller than that in the single-molecule switch because large-area junctions average over many parallel pathways; even in DAS_H- and DAS_Li-based MJs, efficient interfacial pathways are likely retained, but less energetically favorable and with a much smaller population compared to the junctions based on DAS_Na and DAS_K layers, resulting in a medium conductance in these systems.

The importance of ion size and mobility has also been highlighted in viologen-based MJs reported by Han and co-workers. In those devices, the switching ratio increased across the halide series from F^– to I^–, which was attributed to weaker ion binding and higher mobility for larger halides. However, excessively bulky counterions such as ClO₄ ^– and PF₆ ^– degraded the device performance, indicating that ion mobility must be balanced against structural order and leakage suppression. This observation is conceptually consistent with the DAS_X series, in which Na⁺ appears to offer the most favorable balance among ion-pair lability, film compactness, and interfacial accessibility.

Conclusion

To summarize, we demonstrate a robust strategy for incorporating counterions directly into molecular layers during electro-grafting and show that counterion identity has a pronounced impact on charge transport in DAS-based molecular junctions. Compared with postgrafting ion-exchange approaches, , which can disrupt film integrity, this in situ method preserves film compactness and device reproducibility. All DAS-based junctions exhibit a relatively low attenuation factor (β ≈ 1 nm^–1), consistent with a small effective charge-injection barrier produced by a favorable energy-level alignment and ionic screening. Crucially, the conductance difference originates primarily from changes in J ₀ rather than β, indicating that counterions regulate the population and strength of transport-active interfacial coupling pathways. Na⁺ provides the most favorable environment for efficient interfacial coupling, resulting in the highest junction conductance. In contrast, H⁺ and Li⁺ favor more strongly associated ion pairs that limit the formation of efficient contact configurations. These findings identify counterion engineering as a powerful strategy for tuning molecular electronic devices by controlling electrode-molecule coupling at buried interfaces.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.6c08118.

• It includes experimental details on sample preparation (electrochemical grafting); characterization (AFM and XPS analysis) and transport measurement (JV curves for all samples and several thicknesses) (PDF)

The authors declare no competing financial interest.