UK researchers have halved the temperature at which "high-quality" graphene can be grown, claimed the University of Cambridge.
The material can be grown by chemical vapour deposition (CVD) on either nickel or copper catalyst layers.
Key to the Cambridge work is adding a small amount of gold (<1%) to a nickel growth film.
"Once we'd developed a detailed picture of how the graphene was growing, we were able to start tuning that growth and rationally engineering the nickel catalyst to improve it," said researcher Robert Weatherup.
Growth is initiated as a few carbon atoms bond to the metal surface then, as more carbon atoms join the pioneers, these nucleation sites grow sideways into graphene crystals.
Although carbon atoms prefer to attach to an existing carbon island, Weatherup told Electronics Weekly, at lower temperature their mobility on the nickel surface is reduced which makes them more likely to bond to the nickel and start a new crystal than travel around to enlarge an existing crystal.
Until now, he said, it has taken ~1000ºC to grow crystals large enough to build circuits on - around 15µm across.
"Commonly graphene devices around here have 1µm channel widths and are put down in a grid pattern. Larger crystals mean a lower number of devices have a grain boundary through them," said Weatherup.
The gold cuts the chance of nucleation, allowing crystals to grow bigger before they hit one another and form a boundary.
It has reduced the temperature needed to grow 15µm crystals from ~1000ºC to 450ºC - a temperature more compatible with silicon circuits.
Gold also reduces the temperature at which desirable single-atom-thick graphene crystals grow.
According to Weatherup. 74% of the 450ºC nickel-gold substrates are covered with mono-layer graphene.
Adding a small amount of gold to a nickel film reduces the temperature at which usable graphene can be produced from 1,000ºC to 450ºC.
Further work will include trying to eliminate metal completely.
"We would ideally like to produce graphene directly on to an insulating substrate, as at present the alloy has to be removed after growth for graphene to be used in applications," Weatherup said. "The problem is that insulators tend to be less good at converting carbon-containing gases into high-quality graphene."
High quality graphene?
To be useful for semiconductors, graphene should be: monocrystaline, one atom thick, defect-free, and large enough to build devices on.
Ideally, one large crystal would be grown across a conventional silicon wafer for processing in a chip fab.
The best quality graphene is still flaked from lumps of natural graphite which has no grain boundaries, often removed using sticky tape.
Flakes obtained in this way are large crystals, some millimetres across.
One atom thick material is obtained by laboriously refining and sorting the flakes.
'According to University of Cambridge graphene specialist Robert Weatherup, CVD growth can now produce crystals approaching the quality of flaked graphene at the lattice scale.
"There are still defects at the atomic level, but both flaked and CVD are good enough for producing working transistors" he said.'
Better CVD techniques are improving the yield of desirable single-layer graphene and producing larger crystals.
However, crystals are still only a few microns across and, as such, are frequently not big enough to build anything useful on.