Australia explores novel 2D materials for mechanics and electronics

Materials World magazine
7 Aug 2019

The Queensland University of Technology has been exploring organic 2D materials from molecular building blocks.

Synthesis of 2D materials using a bottom-up strategy allows the design of structure at the nanoscale. Building materials on surfaces provides further control of the structure, both through the enforced two-dimensionality due to the presence of the surface, and through the possibility of epitaxy on crystalline surfaces. Suitably chosen molecules deposited on surfaces can spontaneously self-assemble into ordered structures, presenting a rapid and scalable method to produce 2D films and coatings with tailored chemical and physical properties.

The choice of the molecular building blocks and the surface provide access to a near-infinite variety of 2D structures. The surface choice is important – it behaves as a template, with the electronic and structural properties of the surface guiding the positioning of the molecular building blocks. The surface can also modify the deposited molecules, acting as a catalyst, or can react with the molecules, for example by donating adatoms to the structure. Common surfaces used for synthesis are highly oriented pyrolytic graphite (HOPG), graphene, and single crystals of coinage metals. Molecular building blocks are commonly simple organic molecules that have high symmetry and are able to self-assemble through non-covalent interactions.

The study of these 2D layers requires the measurement of physical and chemical properties on the nanoscale. The Surface Science Group is based out of the Nanoscale Imaging Lab, hosted within the Central Analytical Research Facility (CARF) as part of the Institute for Future Environments (IFE) at Queensland University of Technology (QUT), Australia. The IFE offers a range of cutting-edge instrumentation for materials characterisation, with X-ray photoelectron spectroscopy (XPS) and scanning tunnelling microscopy (STM) playing key roles in understanding the form and function of 2D organic materials.

This article will focus on two examples that illustrate how simple organic molecular building blocks can be used to create novel 2D networks on surfaces.

2D substitutional solid solution

Materials that consist of two or more mixed components can offer improved properties over a single phase material. A common example of this is the alloying of metals to produce bronze or steel. To date, self-assembled 2D structures consisting of two or more components have tended to focus on bi-component crystals assembled using complementary molecular building blocks. A typical approach in this field is to engineer complementary building blocks that have matched donor-acceptor hydrogen bonding sites so that the blocks bond into co-crystals. The complementary hydrogen bonding of the co-crystals stabilises the structure, and the material is structurally and stoichiometrically well-defined.

Another possibility for a 2D binary crystal mixture is a randomly intermixed solid solution. In collaboration with groups in China and Canada, this group studied a mixture of two molecular components – trimesic acid (TMA) and terthienobenzenetricarboxylic acid (TTB) on a HOPG surface, revealing a hydrogen-bonded randomly intermixed crystalline phase which has a stable stoichiometry and is thermodynamically stable. This demonstrates the possibility for hydrogen-bonded 2D substitutional solid solutions.

The molecules TMA and TTB both have a planar tripod shape and are C3-symmetric, but are significantly different in size. Both TMA and TTB can self-assemble alone from solution in heptanoic acid to form porous chicken wire structures at the solution/HOPG interface. The structures are stabilised by strong hydrogen bonding between the carboxylic groups that terminate the molecules. The dimensions of the chicken wire structure formed for either the TMA or TTB molecule are a function of their molecular size and their epitaxial match to the HOPG substrate.

TMA forms a network with a hexagonal unit cell with a lattice parameter of 1.61nm and the TTB molecule forms a network with a hexagonal unit cell of 2.37nm. In both lattices, epitaxy induces strain into the network, with the TMA lattice compressed by 3% and the TTB network compressed less than 1% from their expected gas-phase dimensions.

The mixing of TMA and TTB on the HOPG substrate produces a chicken wire network in which TTB randomly substitutes for TMA in 12 ± 1 % of the lattice sites. In the intermixed lattice the pore spacing is measured to be 1.70 ± 0.08nm, and this can be considered the lattice constant of the intermixed system.

Remarkably, the size of this intermixed lattice constant is consistent with Vegard’s law, an empirical rule that states that for a substitutional solid solution a linear relationship exists between the crystal lattice constant (a) and the concentration of the constituent elements or molecules. The position of TTB molecules in the mixed lattice does not exhibit any long range order, consistent with a substitutional solid solution.

The relative stability of the TMA/TTB intermixed phase was examined by evaluating the Gibbs free energy of the intermixed phase relative to the pure TMA phase. Density functional theory calculations were used to understand the enthalpic component for the intermixing, which is made up of intermolecular and molecule-substrate interactions, and Monte Carlo simulations were used to examine the entropy of intermixing. These calculations found that the epitaxy of the intermixed lattice plays a significant role in stabilising the structure, and predicted a minimum in the Gibbs free energy for TTB mixing fractions between 0.10 – 0.12. This suggests that the observed intermixed lattice is thermodynamically stable with respect to the pure TMA phase.

Enthalpy calculations for the intermixing indicate that domain size also plays an important role. As the intermixed domain size increases towards 300 molecules, the calculated enthalpic penalty for substitution also increases. This is possibly associated with an easing of strain on the perimeter of the intermixed domains, an effect that is less significant as the domains become larger and the ratio of perimeter to area reduces. The conclusion from this is that stable intermixed domains are limited in size – between 5-20nm as observed in experiment – and an extended 2D lattice may not be thermodynamically favourable.

This study showed that the observed 2D substitutional solid solution involves a delicate enthalpy/entropy balance where epitaxy, hydrogen bond elasticity, and domain size all have important roles. The stabilisation of this system by epitaxy and the departure of its properties from bulk expectations suggests new engineering pathways for mixed crystals may be available by confinement to a surface. The limitation of size to small nanoscale domains also presents an interesting opportunity, where intrinsic thermodynamic constraints could act to ensure that synthesised alloy lattices are restricted within a bounded range of sizes.

Future experiments on 2D substitutional solid systems should plan to examine other molecular building blocks and determine whether different temperatures can amplify or attenuate the entropy term in the Gibbs free energy. Another possibility is to explore covalently bonded 2D substitutional solid solutions. If a binary mixture of molecular building blocks is self-assembled on surface and annealed to sufficient temperatures reactive groups can be cleaved from the molecules and covalent bonding between molecular units can occur – creating an on-surface 2D polymer. A 2D polymer provides mechanical and thermal stability and the possibility of being removed from the substrate. If the polymer is designed with π-conjugated bonding then it can conduct and has possible electronic and photonic applications. As such, these 2D polymers are of interest in areas such as flexible displays, wearable devices and bioelectronics. A 2D substitutional solid solution polymer would be a heterostructure with intriguing mechanical and electronic properties.

Modern pad design

In addition to providing a predictable hydrogen bonding, TMA’s carboxylic groups (–COOH) also provide access to an additional bonding channel – metal-organic bonding. In solution, the proton can dissociate from the –COOH group, creating a carboxylate group (–COO-) that is eager to form metal coordination bonds. This same deprotonation process can be facilitated on a reactive surface, where the proton can be catalytically cleaved from the –COOH group.

Our group published a paper titled Periodic and nonperiodic chiral self-assembled networks from 1,3,5-benzenetricarboxylic acid on Ag(111), in Chemical Communications, which looked at novel structures of TMA formed following partial deprotonation of its carboxylic groups on a surface.

In this study TMA was deposited onto a single crystal of silver with a (111) orientation, known as Ag(111), where it self-assembled into the same chicken wire hydrogen-bonded network described above. XPS, which is a chemically sensitive probe, confirmed that the TMA molecule was intact and had not reacted with the surface. Further annealing of the surface to 157°C changed the network to a dimer phase that results from partial deprotonation of the carboxyl groups –approximately 33% – and creates a network stabilised by ionic hydrogen-bonds.

Annealing the surface further results in increased deprotonation of the TMA molecules, with a new phase emerging at 200°C, when the TMA is 50% deprotonated. STM imaging of the surface reveals a pinwheel-type structure, which consists of a hexamer of TMA arranged around a bright central protrusion, as the basic building block. The hexamer comprises six TMA molecules, each of which offers a carboxylate group to form a coordination bond to the central bright protrusion, which we interpreted to contain one or more Ag adatoms.

The remaining twelve carboxylic groups around the outside of the pinwheel are intact - not deprotonated. The hexamer possesses an intrinsic chirality which arises due to an approximate 20O rotation of each TMA molecule with respect to the unit vector of the hexagonal cell of the pinwheel network. After further annealing to 220°C, full deprotonation of the TMA molecules is achieved and the observed network structure has evolved. The new network phase, which we term knitted pinwheel, is a condensed version of the previous phase, consistent with complete deprotonation of the TMA and elimination of the hydrogen bonding that previously existed between pinwheels.

The discrete and knitted pinwheel phases described above represent endpoints of a continuum phase that was routinely observed in experiments. The 2D network in between these two deprotonation levels contains a continuous lattice that comprises both discrete and pinwheel domains and lacks long-range order. This phase is essentially a nonperiodic granular alloy of discrete and knitted pinwheels, and has a spatially-varying stoichiometry that depends on the local structure.

In our study, we were unable to determine the number of Ag adatoms at the centre of the pinwheel, but follow up work by K L Svane in the paper An extended chiral surface coordination network based on Ag7-clusters, published in The Journal of Chemical Physics, has determined that seven Ag adatoms form the central bright protrusion of the pinwheel structures, and that the outer atoms pick up a different charge than the inner atom.

The final knitted pinwheel phase consists of a highly ordered array of Ag adatom sites surrounded with a 2D organic framework on top of single crystal Ag. This interesting surface may possibly find applications as a chemical sensor, using the Ag adatoms as detection sites, or as a catalyst for heterogenous reactions, with the Ag atoms acting as reactive sites with a well-defined chemical state.

This study shows that a simple molecule can form the basis for a complex, chiral, 2D self-assembled network. The catalytic properties of the Ag surface were observed as annealing produced several distinct 2D networks, from hydrogen-bonded to fully metal-coordinated and included a non periodic granular alloy phase that contained a mixture of bonding.