Creating a regular array of metal-complexing molecules on an insulator surface at room temperature

Controlling self-assembled nanostructures on bulk insulators at room temperature is crucial towards the fabrication of future molecular devices, e.g., in the field of nanoelectronics, catalysis and sensor applications. However, at temperatures realistic for operation anchoring individual molecules on electrically insulating support surfaces remains a big challenge. Here, we present the formation of an ordered array of single anchored molecules, dimolybdenum tetraacetate, on the (10.4) plane of calcite (CaCO3). Based on our combined study of atomic force microscopy measurements and density functional theory calculations, we show that the molecules neither diffuse nor rotate at room temperature. The strong anchoring is explained by electrostatic interaction of an ideally size-matched molecule. Especially at high coverage, a hard-sphere repulsion of the molecules and the confinement at the calcite surface drives the molecules to form locally ordered arrays, which is conceptually different from attractive linkers as used in metal-organic frameworks. Our work demonstrates that tailoring the molecule-surface interaction opens up the possibility for anchoring individual metal-complexing molecules into ordered arrays.


II. Löwdin Charge Analysis, Charge Density and Charge Displacement Field
We calculated the atomic Löwdin charges of the entire system (molecule adsorbed on the surface), which we compare with the number of valence electrons of the corresponding elements (see Supplementary Table 1    A final, important aspect is the total lack of surface charge redistribution in surface areas already some small distance away from the molecule (see Supplementary Figure 9). This clearly indicates the lack of an electronic screening, potentially affecting the M-M interaction in the presence of the surface.

III. Diffusion Energy Barriers Along the [ ̅ ] and [ ̅ ] Directions Calculated by
Nudged Elastic Band (NEB). We

IV. Diffusion Analysis at 327 K
As shown in this work in Figure 4   at 327 K. At this temperature, the molecule seem to have a clearly higher mobility than at 300 K.
Assuming an attempt frequency of 0 = 10 12 s −1 results again in a diffusion barrier of diff = 1.0 eV and, hence is in perfect agreement with the determined diffusion barrier at room temperature.
This analysis clearly shows that the molecules have a sufficient mobility at 400 K to arrive at their ideal adsorption position during the experiment time.

V. Annealing Experiment
Annealing experiments of MoMo on calcite (10.4) are performed to investigate whether island formation can be induced upon heating. In these experiments, we anneal the sample at a given temperature for one hour, let it cool down to room temperature and image the surface. This procedure is repeated with increasing annealing temperatures to observe the resulting structural changes in the molecular pattern on the surface. As shown in Supplementary Figure 12, AFM images are obtained at room temperature subsequent to annealing to 300 K, 356 K, 492 K, 628 K, 698 K and 738 K. Basically, no change in the random molecule distribution is observed up to an annealing temperature of approximately 700 K. However, above 700 K large cluster are formed, which we tentatively ascribe to a decomposition of the molecules and a clustering of the arising fragments.

VI. DFT Calculations of Intermolecular Interaction
In this Section, we evaluate via DFT the repulsive energy between two MoMo molecules. Our aim is to ascertain the validity of the hard-sphere repulsive model illustrated in the main text and detailed in the Supplementary Discussion, VIII. Hard-Sphere Simulation. In Supplementary   Figure 13 we plot the total energy of the two interacting molecules in gas phase with a fixed orientation, as a function of the distance between their centres along the [421 ̅ ] direction. We adopt the orientation of the most stable adsorption geometry found on the surface (see Figure 3d of the main text). We keep the orientation fixed (the molecule atoms can only move along [421 ̅ ]) to reproduce the constraints imposed by the surface. Based on the charge analysis shown in the Supplementary Discussion, II. Löwdin Charge Analysis, Charge Density and Charge Displacement Field, we can exclude further surface mediated effects like electron screening. We start from a short distance (0.81 nm, experimentally not observed) and compute the total energy until no further change in energy can be detected. The repulsion, vanishing at DFT = 0.91 nm, is related to the electrostatic interaction between the CH 3 groups. Note that when the molecules are allowed to rotate in gas phase (not shown) they manage to relax by keeping a short distance (0.855 nm). However, on the surface this would be energetically unfavourable, as the change in orientation would make the MoMo interaction with the surface Ca atoms far less optimal. As shown in Supplementary Figure 13, the short-range repulsion increases continuously when two MoMo molecules are brought together and not as abruptly as for a rigid sphere. Therefore, in contrast to ideal hard-spheres, MoMo molecules can be slightly compressed under sufficient pressure. However, in a first approximation it remains justified to use a hard-sphere model, because a short-range repulsion only and no intermolecular attraction is recognized.

VII. Non-linear Thermal Drift
In the AFM measurements presented in this work, for example in Figure