Is there a way to make big sheets of pristine graphene or other two-dimensional materials? The answer is blowing in the wind.
That's the heart of a discovery by scientists at Rice University, New Mexico State University and the Department of Energyâ€™s Oak Ridge National Laboratory (ONRL) who grew single-atom-thick graphene monocrystals to unprecedented sizes.
The technique developed by ONRL researcher and lead author Ivan Vlassiouk, New Mexico scientist Sergei Smirnov and Rice materials theorist and co-author Boris Yakobson in principle produces pristine graphene of unlimited size and makes it suitable for roll-to-roll production.
Their process deposits a narrow band of hydrocarbon precursor onto a moving substrate, with a buffer gas blowing the carbon atoms toward the growing front. Once the atoms grab ahold of the substrate and crystallize into a seed of graphene, the buffer wind prompts them to cohere into a single growing sheet.
The researchers reported in Nature Materials their success in growing atom-thin sheets of graphene a foot long and a few inches wide, limited only by the width of the equipment. The single crystal of two-dimensional carbon grows at an inch per hour in a custom-built chemical vapor deposition (CVD) furnace.
The buffering breeze solved a stumbling block for researchers as it helped quash the nucleation of competing graphene seeds on the substrate, which allowed one dominant seed to take control and dictate the growing crystal's orientation. Yakobson's lab modeled how one graphene seed would become the fittest and how it would advance, depending on the substrate and precursors.
This process of evolutionary selection was proposed in 1967 as the mechanism by which 3-D crystals grow via selection of the fastest-growing grains among a random array. In growing silicon ingots to make wafers for microprocessor chips, for example, crystal grains could start growing like forests with a variety of orientations.
"Their growth rates are also different, so some crystals advance faster than others and also become wider," Yakobson said. "Sooner or later, ones that are oriented the same become dominant: They fuse without a grain boundary and form a monocrystal."
He said that's key to pristine 2-D growth as well, but it doesn't come naturally.
When graphene is grown in a typical CVD furnace, crystalline "islands" form on the substrate. They come together as they grow but because they are not turned the same way, carbon atoms adjust where they join to form five- and seven-member rings known as defects. On the larger scale, these appear as grain boundaries that affect graphene's electrical, thermal and optical properties.
The ONRL team solved the problem by building a furnace that pulls the substrate through a thin channel where it is exposed to a two-part stream. The first is a buffer of hydrogen and argon pumped continually through the deposition tube and the second is a hydrocarbon feedstock delivered to the substrate through a small nozzle.
If the conditions are right, only the fittest bit of graphene will be selected. "This is why we also refer to the process as evolutionary," Yakobson said. "It truly is the survival of the fittest crystal. From that point, the crystal can be grown as long as desired.
"My experimental colleaguesâ€™ ingenuity was in suppressing all secondary nucleation," he said. "This is a paradigm shift. From the theoretical perspective, it was compelling to understand which crystal direction wins and how it depends on the catalytic substrate, feedstock and other conditions."
The experimental team found that at the start of deposition, islands did indeed form on the substrate, but after a couple of inches the fastest-growing seed took over and determined orientation going forward.
Making the crystal was one thing; proving they'd done so was another challenge. "In fact, most of the experimental effort went into not the growth itself, but rather analysis of growth outcomes and proving that grown material is monocrystalline," Vlassiouk said.
Because it's impossible to capture an atomic-resolution image of a foot-long crystal, ONRL scientists etched small holes into the graphene and used an automated imager and custom algorithm to build a histogram of the dominant edge angles of the holes.
The histogram revealed three clear peaks showing 60-degree angles to prove the hexagons were consistent throughout, proving the material's global monocrystalline quality. This also revealed the graphene edges were all zigzags, as theory predicted, Yakobson said.
Yakobson noted another advantage: The process does not require a perfect substrate to grow a perfect crystal. "People have tried very hard to get monocrystalline metal for epitaxial growth (in which the orientation of the substrate determines the orientation of crystal growth)," he said. "In this case, the experimental substrate was nothing special. That's a big plus."
The process may simplify the creation of 2-D materials like boron-nitrogen or transition metal dichalcogenides. Yakobson said epitaxial growth of such materials would not quench the formation of grain boundaries, but the new process should eliminate such defects.
He said the main use for a large sheet of graphene would be to cut it into uniform pieces for applications, as with silicon wafers for microprocessors. That way, the orientation of graphene's six-member rings would not matter.
Experiments showed that changing substrates and hydrocarbon precursor also changes the direction of graphene's growth because the catalytic activity is different. Cutting the material along the desired orientation eliminates that issue, Yakobson said.
"If graphene, or any 2-D material, ever rises to industry scale device-making, this method is bound to become a pillar of production to parallel the Czochralski process for silicon," he said.
Co-authors are graduate student Nitant Gupta and research administrator Ksenia Bets of Rice; postdoctoral researcher Yijing Stehle, staff scientist Raymond Unocic, group leader Arthur Baddorf, research scientist Ilia Ivanov, staff scientist Nickolay Lavrik and researcher Frederick List of ONRL; Pushpa Raj Pudasaini, an alumnus of ONRL and the University of Tennessee, and Philip Rack of ONRL and an associate professor at the University of Tennessee. Yakobson is the Karl F. Hasselmann Professor of Materials Science and NanoEngineering and a professor of chemistry.
The Department of Energy and its Basic Energy Sciences division, the Advanced Research Projects Agency-Energy and the Office of Naval Research funded the research.