ARTICLE pubs.acs.org/ac
Electrochemical Fabrication of Highly Stable Redox-Active Nanojunctions Marion Janin, Jalal Ghilane,* Hyacinthe Randriamahazaka, and Jean-Christophe Lacroix* Nanoelectrochemistry Group, Universite Paris Diderot, Sorbonne Paris Cite, ITODYS, UMR 7086 CNRS, 15 rue Jean-Antoine de Ba€if, 75205 Paris Cedex 13, France
bS Supporting Information ABSTRACT: Redox-gated molecular junctions were obtained starting with a relatively large gap between two electrodes, in the micrometer range, followed by electrochemical polymerization of aniline. Polyaniline (PANI) grows from the tip side until it bridges the two electrodes. The resulting junctions were characterized electrochemically by following the variation of the tip substrate current as a function of the electrochemical gate potential for various bias voltages and by recording their I(V) characteristics. The two electrodes make contact through PANI wires, and microjunctions with conductances around 10 3 S were obtained. On the basis of a similar setup, PANI nanojunctions with conductances between 10 7 and 10 8 S were made, where the current appears to be controlled by fewer than 10 oligoaniline strands. Despite the small number of strands connecting the two electrodes, the junctions are highly stable even when several successive potential sweeps are performed. Comparison of the conductance measured in the oxidized and reduced states leads to an on/off ratio of about 70 100, which is higher than that reported for a single aniline heptamer bridging two electrodes, highlighting the interest of connecting a few tens of molecules using the scanning electrochemical microscopy (SECM) configuration. In some cases, the switching of the PANI takes place in several individual conductance steps close to that obtained for a single oligoaniline. Finally, starting with a microjunction and mechanically withdrawing the tip shrinks it down to the nanometer scale and makes it possible to reach the regime where the conductance is controlled by a limited number of strands. This work presents an easy method for making redox-gated nanojunctions and for probing the conductance of a few oligoanilines despite an initially large tip substrate gap.
T
he electrical properties of individual molecules and molecular assemblies are currently of heightened interest because of potential applications in molecular electronics and new opportunities for understanding charge transport in organic systems.1 3 A number of ways have been proposed for measuring conductance through molecules and molecular assemblies, including scanned probe techniques,4 mercury drop electrodes,5 electrical or mechanical break junctions,6,7 and sandwich electrodes.8 The common idea in all of these methods is to “wire” the molecules between two electrodes and to measure the current as a function of an applied bias voltage. In work related to conductance measurement across molecular assemblies, experiments were initially based on making a chemically modified electrode to establish the first electrode molecule contact. Thiol derivatives and electrografting based on diazonium compounds have been used to make thin-film-modified electrodes.9 11 The second electrode is then either brought into contact temporarily, using scanning probe microscopy, or permanently, using crossbar electrodes to measure those molecules that are trapped across the junction. Organic π-conjugated materials are widely used in such studies since they can be switched, by means of various external inputs, between two states with different electronic characteristics. Their electrical doping in field effect transistors is the basis of their r 2011 American Chemical Society
increasing use in low-cost plastic electronics12,13 Electrochemical switching is an easy means of controlling grafted molecules14 and metallic nanoparticle properties.15 They have also been widely used in molecular-scale electronic devices,4,16 18 and their single-oligomer charge-transport properties have been investigated.19 22 Recently, Frisbie’s group investigated the charge-transport mechanism in a conjugated oligo-imine and demonstrated the transition from tunnelling to hopping transport as a function of the length of the entrapped oligomer. 23,24 In this attractive field, electrochemistry has been proposed as an interesting approach for the fabrication and the modulation of metal/molecules/metal junctions.25,26 Electrochemistry makes it possible to change the redox state of the molecule by modulating the potential of the solution relative to the tip and the substrate.27,28 In doing so, the reference electrode plays the role of the gate electrode in solid-state devices,28 but the device has to be immersed in an electrolyte solution. The pioneering work combining singlemolecule investigation and electrochemistry was done by Tao.29 In addition, Lindsay’s group has studied a number of oligomers using the scanning tunneling microscopy (STM)-based break Received: October 24, 2011 Accepted: October 29, 2011 Published: October 30, 2011 9709
dx.doi.org/10.1021/ac202788y | Anal. Chem. 2011, 83, 9709–9714
Analytical Chemistry junction technique under electrochemical potential control.30,31 Similarly, the electrochemical gate-controlled conductance of a single oligomer has been reported.32 34 Despite such successes, connecting a few molecules remains a challenge and there are some discrepancies in the published results. Consequently, finding an ideal technique to study the electrical properties of molecules is still a debate and new methods for connecting a few molecules between two electrodes are still needed. Scanning electrochemical microscopy, SECM, has proved to be a powerful technique for the quantitative investigation of interfacial physicochemical processes.35 Various applications of SECM have been described and detailed in several reviews.36,37 Recently, the development of micro- or even nanofabrication techniques has improved the spatial resolution of SECM by reducing the size of the probe electrode to the nanometer scale.38,39 Even if the SECM configuration has proved useful for studying small numbers of molecules (between 1 and 10) in solution trapped between the tip and the substrate,40 and appears to be well-adapted to setting a gap between the tip and the substrate, the method has not, so far, been used in the fabrication of metal/molecules/metal junctions. In this work we investigate redox-gated π-conjugated molecular junctions. In order to do so, we combine the positioning of two electrodes operating in the SECM configuration and the electropolymerization process. First, a gap was established between the two electrodes, tip and substrate, and then polyaniline was grown electrochemically, starting from the tip side, until it bridged the two electrodes. The resulting junctions were investigated by recording the variation of the current through the polyaniline (PANI) as a function of the potential at fixed bias voltages. Various junctions were generated and investigated, showing the possibility of studying the PANI junction where the conductance is controlled by various numbers of molecules ranging from ∼106 to ∼10 and ultimately to a single oligomer. Finally, by starting with a relatively large PANI junction and moving the SECM tip, the variation of the conductance as a function of the pulling distance was investigated.
’ EXPERIMENTAL SECTION Chemicals. Ferrocenemethanol and ferrocene (Acros) were used as redox couple, and lithium perchlorate (Acros) was used as supporting electrolyte. Aniline monomer (Sigma-Aldrich) was used as received. Sulfuric acid (18 M) was supplied by Prolabo. Electrodes. Disk microelectrodes (UMEs) were made by sealing gold wires (Goodfellow) in soft glass tubes using a laser puller.38 Prior to use, the UME was polished using diamond pastes of decreasing sizes. A quasi-reference electrode was made of Pt wire covered with polypyrrole film.41 Its potential was checked versus the ferrocene/ferrocenium couple (considering E° = 0.400 V/SCE); all potentials were referenced to a saturated calomel electrode (SCE). A platinum wire (1 mm diameter) was used as the counter electrode. Electrochemical Measurements. Electrochemical measurements were performed using a commercial scanning electrochemical microscope (SECM), CHI 900B (CH Instrument, Austin, TX). A four-electrode setup was employed. The UME was used as the working electrode (WE 1), and a 1.5 mm diameter gold electrode was used as the substrate (WE 2).
ARTICLE
Figure 1. Charge-transport current as function of the electrochemical potential (gate potential Vg) for the polyaniline junction using various biases ( 5, 2, 2, and 5 mV). Inset: similarity of the SECM configuration and a field effect transistor.
’ RESULTS PANI Microjunction. As stated above, whatever the method chosen, the study of a molecular junction starts with the positioning of two electrodes separated by a gap. This requirement was achieved by approaching the UME tip (radius 5 μm) close to the substrate using the SECM setup in the feedback mode. When ferrocenemethanol is used as the mediator and the tip approaches the substrate, the tip current increases, indicating that it is approaching an electrochemically active surface. The gap was established by stopping the approach when the tip current increased by approximately 50%, which corresponds to a tip-tosample distance of about 5 μm. The ferrocenemethanol solution was then removed and replaced by a solution containing 0.05 M aniline in 1 M sulfuric acid. Subsequently, aniline was electropolymerized starting from the tip side, the tip potential being held at 0.9 V/SCE. To avoid polymerization at the substrate and spreading of the polymer when the PANI has filled the gap, its potential was set at 0.3 V/SCE. Supporting Information Figure S1 shows details on the PANI growth process. A published procedure42 was used to estimate PANI thickness. The film is found to be around 2 μm thick, a value smaller than the initial tip substrate gap, 5 μm, which suggests a dendritic growth from the tip to the substrate. The contact between the two electrodes through the PANI junction was confirmed and characterized by transport measurements under electrochemical control. Electrochemistry was used to modulate the junction conductance by altering the charge state of the redox system. The inset in Figure 1 shows the similarity of the SECM setup, when PANI bridges the two electrodes, and the classical setup used in field effect transistors. In such a configuration the SECM tip and the substrate play the role of drain and source, respectively, while the reference electrode is equivalent to the gate. As for a solid-state transistor, the variation of the source drain (substrate tip) current versus the gate (tip reference) potential was recorded. The tip and substrate potentials are scanned Etip = VSD) is simultaneously, while a fixed bias (Esubstrate maintained between the two electrodes. Figure 1 shows the variation of the source drain current versus the electrochemical potential (Etip Eref = Vg gate potential) at different VSD biases. The current starts to increase at around 0.15 V, where PANI starts to be oxidized, and reaches a maximum at 0.4 V. Beyond 9710
dx.doi.org/10.1021/ac202788y |Anal. Chem. 2011, 83, 9709–9714
Analytical Chemistry this potential, the current starts to fall as the PANI is further oxidized. Such behavior confirms that the source and drain are connected through the PANI. In the absence of contact no current flows between the two electrodes, and only an electrochemical current in the nanoamp range, corresponding to the response of PANI, is observed at the tip. The shape of this curve is similar to that reported for a junction obtained by the STM setup.26,30,43 The polyaniline oxidation states range from fully reduced (insulating below 0.1 V), to partially oxidized (highly conducting between 0.15 and 0.6 V), and to fully oxidized (insulating and unstable, generally observed above 0.6 V/SCE under such conditions).44 In the present case, the conductivity starts to decrease at 0.4 V/SCE, which may be related to the formation of highly doped nanodomains of low conductivity. During the backward scan the PANI junction is reversibly switched from the conducting to the insulating state. However, the reversibility shows a hysteresis induced by structural relaxation of the junction during the scan. This behavior is fully reversible, and the microjunction was electrochemically characterized several times by applying different bias potentials and changing the direction of bias, as shown in Figure 1. Depending on the polarity of the applied bias, the current flows from the tip to the substrate (positive bias) or from the substrate to the tip (negative bias). Moreover, the change of the current direction through the junction does not affect the shape of the curve. Indeed, the data are perfectly symmetrical when positive and negative biases are compared. Another interesting observation is the sensitivity of the PANI junction to small bias variations. Under a 0 mV bias voltage, the current versus gate voltage curve is that of PANI electroactivity, as shown in Supporting Information Figure S1B, with the current reaching 50 nA and corresponding to doping and dedoping processes. A bias of (1 mV is enough to make the current through the PANI junction reach a few microamps. From Figure 1 the microjunction conductance, in the oxidized state at 0.4 V/SCE, is estimated to be about 10 3 S, leading to a PANI conductivity of at least 1 S 3 cm 1. This value represents the minimum conductivity of the PANI, because in the calculation the section was assumed to be that of the UME tip section. However, the growth of PANI using this procedure is likely to be dendritic and the real section connecting the substrate is smaller than the tip section. Under our conditions the formation of nanowires of polyaniline near the substrate is thus expected, and the conductivity of such nanowires was found to be between 10 and 100 S 3 cm 1.45 These first results demonstrate the possibility of using an SECM as the starting configuration for the fabrication of a highly stable redox-gated conducting polymeric junction. The conductance estimated to be 10 3 S at 0.4 V indicates that many PANI wires connect the tip and the substrate, and therefore, such a junction can be considered as large. Nanojunctions. A number of junctions were created and characterized, with the aim of reducing the amount of PANI in contact with the substrate electrode. In the following, a junction using a 1 μm radius UME and a tip substrate gap close to 1 μm will be presented. Supporting Information Figure S2A shows the variation of the UME current versus time during polymerization, while Supporting Information Figure S2B shows the response of the PANI film deposited. The estimated thickness of the polymer is, in the present case, around 150 nm, which is much smaller than the initial tip substrate gap (this tendency was found to be general in all experiments). However, since PANI reaches the
ARTICLE
Figure 2. Current through the polyaniline junction as a function of electrochemical potential (gate voltage Vg) using various biases: 20, 5, 5, and 20 mV.
substrate, dendritic growth rather than uniform growth occurs, which indicates that the section of PANI wires connecting the substrate is much smaller than that of the tip and reaches the nanometer scale. The PANI junction was electrochemically characterized; Figure 2 shows the variation of the source drain current versus the gate voltage at different fixed biases between the tip and the substrate. Below 0.15 V, the junction is in an insulating state and the current is probably essentially due to leakage (
