Scientists take steps to grow large-scale sheets of 2D insulator

A scientist at Rice University, together with collaborators in Taiwan and China, has successfully grown atom-thick sheets of hexagonal boron nitride (hBN) as two-inch diameter crystals across a wafer. The scientists report their success in a paper in Nature.

Surprisingly, they achieved the long-sought goal of making perfectly ordered crystals of hBN, a wide band gap semiconductor, by taking advantage of disorder among the meandering steps on a copper substrate. These random steps keep the hBN in line.

Set into chips as a dielectric between layers of nanoscale transistors, wafer-scale hBN should excel in damping down the electron scattering and trapping that limit the efficiency of an integrated circuit. Until now, however, nobody has been able to make perfectly ordered hBN crystals that are large enough to be useful.

Materials theorist Boris Yakobson at Rice University’s Brown School of Engineering is co-lead scientist on the study, together with Lain-Jong (Lance) Li of the Taiwan Semiconductor Manufacturing Co. (TSMC) and his team. Yakobson and Chih-Piao Chuu of TSMC performed theoretical analyses and first principles calculations to unravel the mechanisms of what their co-authors saw in their experiments.

As a proof of concept for manufacturing, experimentalists at TSMC and Taiwan's National Chiao Tung University grew a two-dimensional (2D) hBN film with a two-inch diameter, transferred it to silicon and then placed a layer of field-effect transistors patterned onto 2D molybdenum disulfide on top. "The main discovery in this work is that a monocrystal across a wafer can be achieved, and then they can move it," Yakobson said. "Then they can make devices."

"There is no existing method that can produce hBN monolayer dielectrics with extremely high reproducibility on a wafer, which is necessary for the electronics industry," Li added. "This paper reveals the scientific reasons why we can achieve this."

Yakobson hopes the technique may also apply, with some trial and error, to other 2D materials. "I think the underlying physics is pretty general," he said. "Boron nitride is a big-deal material for dielectrics, but many desirable 2D materials, like the 50 or so transition metal dichalcogenides, have the same issues with growth and transfer, and may benefit from what we discovered."

In 1975, Intel's Gordon Moore predicted that the number of transistors on an integrated circuit would double every two years. But as integrated circuit architectures get smaller, with circuit lines down to a few nanometers, that pace of progress has been hard to maintain.

The ability to stack 2D layers, each with millions of transistors, may overcome such limitations if they can be isolated from one other. Insulating hBN is a prime candidate for that purpose because of its wide band gap.

Despite having ‘hexagonal’ in its name, monolayers of hBN as seen from above appear as a superposition of two distinct triangular lattices of boron and nitrogen atoms. For the material to perform as required, hBN crystals must be perfect; that is, the triangles have to be connected and all point in the same direction. Non-perfect crystals have grain boundaries that degrade the material's electronic properties.

For hBN to become perfect, its atoms have to precisely align with those of the substrate it grows on. The researchers found that copper in a (111) arrangement – the number refers to how the crystal surface is oriented – does the job, but only after the copper is annealed at high temperature on a sapphire substrate and in the presence of hydrogen.

Annealing eliminates grain boundaries in the copper, leaving a single crystal. Such a perfect surface would, however, be ‘way too smooth’ to enforce the hBN orientation, Yakobson said.

Last year, Yakobson reported on his work growing pristine borophene on silver (111) (see Silver helps borophene grow to unprecedented lengths), and also on a theoretical prediction that copper can align hBN by virtue of the complementary steps on its surface (see Small steps to synthesizing seamless 2D crystals). The copper surface was vicinal – that is, slightly miscut to expose atomic steps between the expansive terraces. That paper caught the attention of industrial researchers in Taiwan, who approached the professor after a talk there last year.

"They said, 'We read your paper'," Yakobson recalled. "'We see something strange in our experiments. Can we talk?' That's how it started."

Informed by his earlier experience, Yakobson suggested that thermal fluctuations allow copper (111) to retain step-like terraces across its surface, even when its own grain boundaries are eliminated. The atoms in these meandering steps present just the right interfacial energies to bind and constrain hBN, which then grows in one direction as it attaches to the copper plane via weak van der Waals forces.

"Every surface has steps, but in the prior work, the steps were on a hard-engineered vicinal surface, which means they all go down, or all up," Yakobson said. "But on copper (111), the steps are up and down, by just an atom or two randomly, offered by the fundamental thermodynamics."

Because of the copper's orientation, the horizontal atomic planes are offset by a fraction to the lattice underneath. "The surface step-edges look the same, but they're not exact mirror-twins," Yakobson explained. "There's a larger overlap with the layer below on one side than on the opposite."

That makes the binding energies on each side of the copper plateau different by a tiny 0.23 electron volts (per every quarter-nanometer of contact), which is enough to force docking hBN nuclei to grow in the same direction, he said. The experimental team found the optimal copper thickness was 500nm, which is enough to prevent its evaporation during hBN growth via chemical vapor deposition of ammonia borane on a copper (111)/sapphire substrate.

This story is adapted from material from Rice University, with editorial changes made by Materials Today. The views expressed in this article do not necessarily represent those of Elsevier. Link to original source.