Single-molecule transport at a rectifying GaAs contact.

In most single- or few-molecule devices, the contact electrodes are simple ohmic resistors. Here we describe a new type of single-molecule device in which metal and semiconductor contact electrodes impart a function, namely current rectification, which is then modified by a molecule bridging the gap. We study junctions with the structure Au STM tip/X/n-GaAs substrate, where ‘X’ is either a simple alkanedithiol or a conjugated unit bearing thiol/methylthiol contacts, and we detect current jumps corresponding to the attachment and detachment of single molecules. From the magnitudes of the current jumps we can deduce values for the conductance decay constant with molecule length that agree well with values determined from Au/molecule/Au junctions. The ability to impart functionality to a single-molecule device through the properties of the contacts as well as through the properties of the molecule represents a significant extension of the single-molecule electronics ‘tool-box’.


Main Text
As the density of components on integrated circuits has increased, their dimensions have decreased to just a few nm. 1 At this length-scale, individual organic molecules represent an exceptionally attractive class of component for electronic devices. They are mechanically flexible, insensitive to the manufacturing defects that can affect components produced by conventional top-down lithography, and add functionality such as the ability to respond to environmental stimuli. These can include illumination, 2-5 temperature changes [6][7][8] and the presence of specific molecules in their surroundings, [9][10][11][12][13] which makes them of special interest for sensor applications. Hybrid organic molecule-inorganic semiconductor devices have the significant advantage that their manufacture could take advantage of the substantial infrastructure investment in the conventional semiconductor industry (where annual expenditure exceeds $40 billion 14 ). However, their development will require a better understanding of the electronic properties of individual molecule-semiconductor interfaces. Although the technological deployment of single molecular junctions would appear to lie a long way off, due to challenges with up-scaling and large-scale fabrication and junction stability, fundamental work is greatly contributing to the understanding of charge transport in molecular junctions and how electrical functionality can be controlled in such junctions. However, combining the properties of semiconducting junctions and single molecules to achieve new functionality has been barely explored and this provides the perspective for the present study.
Organic layers have been fabricated previously on GaAs, as a way to modify the electronic properties of this technologically important surface. For example, organic layers can cause substantial reduction of the surface recombination velocity (SRV) 15 and unpin the surface Fermi level. 16 Furthermore, high 5 quality thiol monolayers on GaAs 17 provide good passivation against oxidation of the surface to Ga2O3 / As2O3, so that samples are stable for days with minimal increase of XPS Ga 3d5/2 and As 3d5/2 oxide signals. 18 Contact angle measurements confirmed the stability of such monolayers for longer periods of time. 19 The electrical properties of relatively large area (1 -10 µm 2 ) Metal / Molecule/ Semiconductor (MMS) devices have been measured previously and are reported extensively in the literature. [20][21][22][23][24] Devices were mainly prepared by depositing a metallic contact on top of a pre-assembled monolayer on a semiconductor, and I-V (current-voltage) characteristics are measured to obtain insights into the mechanism of charge transport and device behavior. The deposition of the metallic contact on top of the monolayer is a significant challenge in these devices, as conventional methods such as sputtering or vacuum evaporation can (and often will) damage the molecules or substantially modify the surface in an uncontrollable fashion. 20,25,26 Therefore, alternative methods for the fabrication of MMS devices have been developed, such as soft evaporation 22 and lift-off, float-on (LOFO) of pre-formed metallic pads. 25 However, in all these methods the presence of pinholes in the monolayer can introduce short circuits in the device, thus making the measurements of its properties unreliable.
To overcome these problems, Lee et al. used a different approach to measure the properties of a MMS device consisting of a GaAs/dithiol monolayer decorated with Au nanoparticles. By contacting a nanoparticle with a STM probe they were able to collect I-V curves and characterize the metal/dithiol/GaAs junction. 23 This method, while advantageous in terms of sample preparation, still has its shortcomings. In particular, the Au nanoparticle needs a stabilizing monolayer on its surface to avoid clustering and sintering, which can introduce additional potential barriers in the device. We reasoned that it would be possible to simply use a Au STM tip to contact the monolayer, thus removing 6 the need for a nanoparticle and reducing the MMS device complexity. This procedure does indeed generate viable Schottky diodes consisting of an Au STM tip and a GaAs substrate coupled by a small number of molecules connected in parallel. The combination of metal and semiconductor contacts are responsible for the rectifying behavior of our device, as in a conventional metal-insulatorsemiconductor diode, while the actual rectification ratio depends on the choice of molecule. Although rectification is a widely studied phenomenon, in both large area and single molecule devices, [26][27][28][29][30][31] we emphasize here that the rectification in the present case requires both the semiconductor contact and molecule. Furthermore, the current through the device is sensitive to the number of molecules connecting it even though the potential difference across the molecules is only a small fraction of the potential across the complete device, and we can detect the attachment and detachment of individual molecules in the device, thus permitting the measurement of properties of single-molecules for the first time in a MMS device.
We focused on three families of molecular wires: (i) simple alkanedithiols (4DT -7DT), (ii) a methylthiol-terminated phenyl ring 1[Ph]1, where the π-system is decoupled from the contacts, and (iii) a fully-conjugated 4,4'-dithiol-1,1'-biphenyl BPDT. Alkanedithiols on GaAs have already been studied in the literature in conventional (large area) MMS devices 21,22,32 and as uncapped monolayers, 18,19,33 and were found to form densely-packed, defect-free layers and devices. As an archetypal molecular electronics system, 34-36 they were the first subject of our study. In a typical experiment, a gallium-indium eutectic ohmic contact is annealed on the back of a n-type Si-doped <100> GaAs wafer (doping density 3 x 10 18 cm -3 ) at 400 ºC in vacuum (∼10 -2 mbar) for 90 minutes. The wafer is then chemically etched (NH4OH 30% in H2O for 5 minutes, followed by DI water rinse) to remove gallium and arsenic native oxides, and then immediately immersed in a degassed ethanol solution containing 1 mM of the desired molecular wire and 5% NH4OH (to avoid oxide layer regrowth and deprotect the thioacetate function in the case of 7DT and BPDT). 37,38 Samples were incubated under Ar atmosphere for 24 hours, removed from solution, thoroughly rinsed with ethanol, dried under a stream of Ar and placed on a Au slide (Arrandee gold-on-glass) with an 8 additional layer of GaIn eutectic painted to provide optimal contact (schematics of device structure in Figure 1b). We started our investigation by analyzing the time-dependent tunneling current for 1,6-hexanedithiol (6DT) monolayers on GaAs as a function of STM setpoint current I0. With a low initial I0 the STM tip is outside the monolayer and the tunneling current profile is flat (green traces, Figure 2a). As I0 is gradually increased and the electrode separation thereby reduced, the tip comes sufficiently close to the monolayer to interact with it, and Au-S bonds spontaneously form. This results in changes of the tunneling current as the molecular bridges are formed, which we observed as sudden jumps in the We can model the devices qualitatively as leaky Schottky diodes, with a thin insulating layer between the metal and the semiconductor (the organic monolayer) and surface states at an energy that is fixed relative to the semiconductor band edges, as shown in Figure 4. In forward bias, a relatively large current is attributed to tunneling from to the metal through the molecular bridge by a mechanism similar to that operative in Au-molecule-Au junctions (for the molecules studied here, we expect transport to be dominated by the HOMO in forward bias, vide infra), but a substantial limitation to the current is the barrier height provided by the conduction band bending of the semiconductor (Figure 4).
In reverse bias, however, electrons can easily tunnel from the metal to by the same mechanism, but they will be trapped there by the conduction band potential barrier. As the reverse bias voltage is increased, the Fermi level of the metal will approach closer to the molecular LUMO, and the number of electrons tunneling through the molecular orbital directly into the conductance band of the semiconductor (green path on Figure 4) could increase. Therefore, the energy of the LUMO relative to the metal Fermi level is likely key to controlling the current in reverse bias. Molecules with a large HOMO-LUMO gap such as alkanedithiols will have the LUMO far in energy from the metal Fermi level, so that the current in reverse bias will be small, and the device will show a high rectification ratio. As the molecular bridge is made more conjugated in nature, the HOMO-LUMO gap will reduce in size, and the LUMO be easier to access under reverse bias, thus allowing a larger charge flow and reducing the rectification ratio. A full transport DFT study of the devices presented is outside the scope of this paper, but as a first approach we can use the calculations performed on molecules sandwiched between metal electrodes to estimate the position of the LUMO relative to the Fermi energy of the metal ( Figure 5). The position of the LUMO with respect to the metal EF matches the order of decreasing rectification observed in the MMS devices presented here: the further is the LUMO from the metal EF, (and thus the bigger is the bandgap) the higher is the rectification ratio.  are observed, and these jumps have been related to a change in charge transport from tunneling through air to tunneling through the molecular backbone. [39][40][41] A typical 500 ms trace contains 3 -8 current jumps. Between each trace the feedback loop was turned on to ensure consistent substrate-tip separation throughout the measurements. The STM setup was kept in the dark for the whole duration of the measurements to avoid the generation of a photocurrent. Hundreds of current jumps were collected this way over several hours, and processed using software written in Python which is described in the SI. Automated algorithms are commonly used to process data in single-molecule electronics measurements. 30,51,52 In our software, the background setpoint current was determined and then subtracted from the raw current vs. time traces which were afterwards compiled into histograms.
Individual traces were broken into segments by locating jumps between the different current levels using features in the differential of the current (dI/dt). This was used to produce the current vs time density plots (Figures S7 to S10).

Associated Content
Details on data analysis, supporting results (current histograms and traces) and current vs time analysis.
This material is available free of charge via the internet at http://pubs.acs.org. Raw data is available on the catalog in Liverpool at: http://datacat.liverpool.ac.uk/133/