Kinetics drive formation of Cu islands beneath graphite surfaces
Earlier this year, a team of experimentalists and theorists at the Ames Laboratory and Iowa State University discovered that multilayer copper islands can form beneath the topmost layers of graphene in bulk graphite. As reported in a recent issue of The Journal of Physical Chemistry C, they have now determined the optimal formation conditions and characterized the morphology of the islands. The results show that island formation is driven by kinetics, not thermodynamics.
Inserting guest atoms or molecules between the layers of a bulk two-dimensional (2D) material is a general, robust way to tune the properties of the material. This technique, bulk intercalation, has been well-studied in graphite but less is known about surface intercalation, which could be an efficient way to tune the surface properties of graphite.
The observation of copper (Cu) islands intercalated in graphite is especially interesting because elemental transition metals like copper do not form bulk intercalation compounds with graphite. In addition, graphene-copper systems display characteristics highly relevant for technological applications. For example, sandwiching Cu between sheets of graphene prevents Cu from oxidation and increases its thermal conductivity.
To induce surface intercalation of Cu, Ann Lii-Rosales, a student working with Patricia Thiel at Ames Laboratory and Iowa State, bombarded a graphite surface with argon ions, used physical vapor deposition to deposit Cu at different temperatures of the substrate, and characterized the samples with scanning tunneling microscopy (STM) and x-ray photoelectron spectroscopy (XPS). The experiments took place in an ultrahigh vacuum chamber.
The results showed that Cu structures evolved with deposition temperature. Below 600 K, STM images revealed rough Cu clusters on the surface of the graphite and no intercalation. But abruptly at 600 K, smooth, faceted structures with sloping sides emerged. The imaging, combined with XPS data, showed that the structures were multilayered metallic Cu islands encapsulated by a graphitic lattice.
The growth of these islands competed not only with the growth of adsorbed Cu clusters at low temperature, but also with small pancake-like features that developed at 850 K and above. The optimal temperature for the formation of large encapsulated islands was 800 K. At that temperature, islands were up to several hundred nanometers wide and 2 nm to 40 nm high. This height range corresponds to 7 to 200 Cu layers.
The data revealed that multilayer islands can have round or flat tops, can be covered by multiple graphene layers, and are stable when exposed to air. Images also showed that the large islands only formed in samples first bombarded with ions, which created plentiful defects.
“There were many surprises in this work,” Thiel says. “One was that the intercalated Cu islands are simply so large and abundant. That observation built an expectation in our minds that this was a thermodynamically driven phenomenon.”
To test this idea, theorist Yong Han compared the stability of adsorbed and intercalated Cu atoms, clusters, and layers in different configurations using density functional theory (DFT). Surprisingly, he found that intercalation is only thermodynamically favorable for single Cu atoms. This implies that island formation is driven by a kinetic effect, in which single atoms intercalate and become trapped.
The researchers theorize that the ion-induced defects play a central role in this process. At 600 K and above, Cu atoms can pass through large defects with relative ease. Once below the surface, kinetics drives the nucleation of multilayer islands, or, if the temperature is above 800 K, of pancake-like single layer clusters. The researchers are currently developing a quantitative model of this process to compare with experimental data. They are also using theory to further investigate the formation and role of the defects.
“Our work opens the road to a new type of Cu-graphite matrix morphology,” Thiel says. In addition, the results may be applicable to other materials systems as well. “It is quite likely that these results can be extrapolated to multilayer-supported graphene, and it is possible that they can be extrapolated to other 2D materials such as bismuth(III) selenide, which are currently of great interest as quantum materials,” she says.
“Scientifically, this is a very thorough work combining both theory and experiment,” says Edward Conrad, an expert in surface physics from the Georgia Institute of Technology who was not associated with this research. He notes that this research has the potential to improve heat transfer devices, given the ability to control the density of copper, prevent it from oxidizing, and achieve intimate contact between copper and graphene. For example, silicon carbide devices could be coated with this Cu-graphene interface for more efficient cooling in high voltage, high power applications, he says.
Read the abstract in The Journal of Physical Chemistry C.