Building Organometallic Chains Molecule by Molcule
- Fig.1: Illustration of a break junction for measurements in liquid environment. The typical length of free-standing gold bridge (u) is 300-500 nm, and the width of its constriction is 60-100 nm. Figure adapted from .
- Fig.2: Conductance-displacement histogram of BDNC-molecules in solution. Three areas of high counts as well as three peaks on the conductance histogram (right) are the signatures of three con-secutive plateaus in conductance traces which are referred to stepwise chain formation. The inset shows the situation when the third plateau is not formed. Figure reproduced from .
- Fig.3: DFT-MD calculations demonstrating the chaining of two BdNC molecules while the gold elec-trodes are withdrawing. Figure reproduced from .
The concept of single-molecule electronics – the use of individual organic molecules as active elements in electrical circuits – strongly roots in the inspirational theoretical contribution by Aviram and Ratner where a molecular diode was proposed . We describe here recent experimental progress showing how an organometallic chain can be assembled molecule by molecule between metallic electrodes.
The idea of using molecules as electronic components is in line with a general tendency followed by modern electronics, the miniaturization of individual components. The great promise of organic molecules over existing semiconductor electronic elements are: 1) their size – most of the simple molecules are significantly smaller than existing individual transistors, 2) their uniformity – molecules of the same compound are identical to the last atom, and 3) the flexibility in design – chemistry provides various well-established routes to tune the properties of molecules. These prospects paved the road for further investigations in the field.
Creating Electrodes on the Atomic Scale
A very successful technique to characterize electrical transport at the nanoscale are mechanically controlled break junctions (MCBJ). Historically, they were developed to study the properties of weak links in superconducting materials using freshly broken metallic surfaces. The technology was then adapted for the study of atomically-sharp metallic electrodes where the quantization of the electrical conductance can be observed, and further on for the measurements of molecules [2, 3]. Nowadays, the MCBJ approach is one of the most common experimental techniques in the field of molecular electronics. The technique relies on the breaking of a metallic wire by pulling it in a controlled way. This is achieved by bending a flexible substrate on which a metallic constriction has been fabricated. Using modern nanofabrication techniques, metallic nanobridges with an attenuation factor (the ratio between the wire elongation and the substrate bending) below 10-4 can be achieved. In other words, bending the substrate by 1 micrometer leads to a wire elongation of less than 0.1 nanometer, resulting in remarkable mechanical stability and tunability of the system.
The breaking of the metallic wire is monitored by applying a potential difference between its two ends and measuring the resulting current in the circuit (Fig.1). The magnitude of the current decreases as the metal wire is stretched until it ultimately consists of a single atom. Upon bending the substrate even further, the nanobridge is broken, accompanied by a drastic change in current as the resulting tunneling gap between the two atomically sharp electrodes slowly increases. To perform the electrical characterization of molecules, the latter are functionalized with ‘anchor groups’ – chemical groups that provide covalent or van der Waals binding to the electrodes and allow molecules to bridge the gap between them. Typical anchor groups are, for instance, thiols (-SH), amines (-NH2), cyanides (-CN) and pyridils (-C5H5N), with gold being the predominantly used electrode material. The properties of molecular junctions formed in this way are defined by: 1) their chemical backbone, 2) the anchor groups, 3) the nature of the electrodes, and 4) the environment (e.g. a solvent).
During the measurement process the junction is repeatedly broken (opened) and then reformed (closed) in the presence of molecules in solution. For every cycle, the change in conductance with electrode separation (conductance trace) is recorded. One or few of the molecules from the surrounding medium can bridge the gap between the electrodes, and this event is reflected in the conductance trace. In this way, the binding and unbinding of a single molecule in a nanometer-sized gap can be detected in real time. To capture the large amount of possible conformations of the molecule(s) between the electrodes, this process is repeated hundreds of times, and then statistically analyzed.
The data are usually presented as either conductance histograms or conductance-displacement histograms (Fig.2). In a conductance histogram, the probability-distribution of conductance values during the entire breaking process is presented. If the target molecule forms stable molecular junctions, a peak appears in the conductance histogram. For the conductance-displacement histogram, a two-dimensional distribution of conductance values as a function of electrode displacement (roughly the gap size) is presented. In this case, the signature of molecular junction formation is a conductance plateau (i.e. a relatively flat area of high counts).
Knitting with Molecules
Because future applications based on molecular junctions require their stable integration into electronic circuits, studying the interaction of molecules with electrodes is of particular interest. In a recent study, we used the MCBJ approach to characterize BdNC molecules . These are benzene molecules that have highly polar isocyano (-NC) chemical groups as anchors. The interaction of isocyanides with gold is known to be particularly strong, with calculations confirming their binding energy to surpass even that of covalent sulfur-gold bonds. Moreover, the isocyano-gold bond is highly directional allowing for dense molecular surface coverage which makes these molecules an ideal test system for reliable junction formation.
The measurements were performed on a 100 micromolar solution of BdNC molecules in a mixture of THF and mesitylene. Multiple conductance plateaus were observed in the conductance histograms of opening-traces with a very high yield of plateau formation (Fig.2). This can be explained by the formation of multiple stable molecular configurations during the measurements process. In notable contrast to break junction measurements on most other molecules, a similar behavior was also observed for the closing conductance traces, i.e. while the electrodes are approaching. This is a result of the polarity of the isocyano-gold bond resulting in molecules sticking out from the surface, ready to form a contact.
To attribute the conductance signatures to specific molecular configurations we performed theoretical calculations using Density Functional Theory based Molecular Dynamics (DFT-MD) simulations at room temperature.
These allow us to visualize the interaction of molecules while pulling apart the electrodes at a temporal resolution of 1 fs. Evaluating several sets of these computational experiments, we are able to link molecule-electrode configurations to observed conductance features. The first conductance plateau with higher conductance was attributed to the single-molecular junction and its conductance value is similar to that observed in other molecules of comparable structure. The second, lower conductance plateau, corresponds to the configuration of organometallic chains which include two molecules plus one or more additional gold atoms. These chains form during the opening process because the strong interaction between the isocyano anchor-group and gold allows the BdNC molecule to pull a gold atom from the electrode. The resulting organometallic compound can then interact with another molecule which is present in close proximity to the junction as a consequence of the dense surface packing described above. Detailed analysis reveals that in 29 % of all breaking traces a third plateau is observed, which is attributed to the formation of trimer chains at even larger electrode displacement and with correspondingly lower conductance. Even longer chains are expected to form, however, with conductance below the detection limit.
As the on-surface concentration of molecules in the vicinity of the initial mono-molecular junction is of prime importance for the chaining process, we employ two complementary approaches to exercise control over organometallic chain formation. First, the same measurements were performed in solutions with smaller concentrations of BdNC molecules to reduce the equilibrium density of molecules on the electrode surface. In the second approach, the central benzene ring of the molecule was extended by two bulky side groups, namely methyl and tert-butyl instead of the usual hydrogen termination. These side groups are expected to decrease the effective electrode surface coverage through steric hindrance, and hence suppress the formation of dimers and trimers. Indeed, experimentally both routes were tested and yielded a suppression of the oligomerization process. Instead, a single high-conductance plateau with modified shape (width and slope) was observed and attributed to a slightly modified configuration statistics of the mono-molecular junction.
The formation of molecular dimers and trimers through incorporation of gold atoms can be considered as a controlled step-by-step synthesis of such a conductive metal-organic 1D-oligomer. Here, the exceptional stability of the MCBJ technique allows for the real-time observation of the process of chain formation, molecule after molecule, as revealed by their conductance properties. In stark contrast to other synthetic schemes, the process does not rely on the stepwise addition of different constituents of the organometallic compound, but occurs via in-situ extraction of atoms from the electrode by the isocyano-molecules. These findings pave the way for the controlled formation of one-dimensional, single coordination chains, which may be used as promising building blocks for organometallic frameworks.
Anton Vladyka1, Jan Overbeck1, Mickael Perrin1, Michel Calame1
1Transport at Nanoscale Interfaces, Empa, Dübendorf, Switzerland
Head of Laboratory
Transport at Nanoscale Interfaces
 Aviram, A.; Ratner, Molecular rectifiers, M. A. Chemical Physics Letters 1974, 29 (2), 277–283, DOI:10.1016/0009-2614(74)85031-1.
 Agraït, N.; Yeyati, A. L.; van Ruitenbeek, Quantum properties of atomic-sized conductors, Jan M. Physics Reports 2003, 377 (2-3), 81–279, DOI:10.1016/S0370-1573(02)00633-6.
 See e.g. Focus issue on Molecular electronics, Nat. Nano. 2013, 8 (6), 377-467 (2013).
 Vladyka, A.; Perrin, M. L.; Overbeck, J.; Ferradás, R. R.; García-Suárez, V.; Gantenbein, M.; Brunner, J.; Mayor, M.; Ferrer, J.; Calame, In-situ formation of one-dimensional coordination polymers in molecular junctions, M. Nature Communications 2019, 10 (1), 262, DOI:10.1038/s41467-018-08025-9.