This is the first demonstration of an industrial, scalable method for manufacturing high-quality graphene that is tailored for use in membranes that filter a variety of molecules, including salts, larger ions, proteins, or nanoparticles.
“For several years, researchers have thought of graphene as a potential route to ultrathin membranes,” says John Hart, associate professor of mechanical engineering and director of the Laboratory for Manufacturing and Productivity at MIT. “We believe this is the first study that has tailored the manufacturing of graphene toward membrane applications, which require the graphene to be seamless, cover the substrate fully, and be of high quality.”
Graphene is often synthesised through chemical vapour deposition, heating a copper foil that acts as a substrate for a combination of carbon and other gases. If graphene membranes are ever to be used commercially they will have to be produced in large quantities, at high rates, and with reliable performance.
“We know that for industrialization, it would need to be a continuous process,” said Hart. “You would never be able to make enough by making just pieces. And membranes that are used commercially need to be fairly big – some so big that you would have to send a poster-wide sheet of foil into a furnace to make a membrane.”
The team used a roll-to-roll approach with CVD to manufacture high-quality graphene in large quantities and at a high rate. The system consists of two spools, connected by a conveyor belt that runs through a small furnace. The first spool unfurls a long strip of copper foil, less than 1cm wide. When it enters the furnace, the foil is fed through first one tube and then another, in a “split-zone” design.
While the foil rolls through the first tube, it heats up to a certain ideal temperature, at which point it is ready to roll through the second tube, where the scientists pump in a specified ratio of methane and hydrogen gas, which are deposited onto the heated foil to produce graphene.
“Graphene starts forming in little islands, and then those islands grow together to form a continuous sheet,” said Hart. “By the time it’s out of the oven, the graphene should be fully covering the foil in one layer, kind of like a continuous bed of pizza.”
As the graphene exits the furnace, it rolls onto the second spool. The researchers found that they were able to feed the foil continuously through the system, producing high-quality graphene at a rate of 5cm per minute. Their longest run lasted almost four hours, during which they produced about 10m of continuous graphene.
Once the researchers produced graphene using their roll-to-roll method, they unwound the foil from the second spool and cut small samples out. They cast the samples with a polymer mesh, or support and etched away the underlying copper.
“If you don’t support graphene adequately, it will just curl up on itself,” Kidambi says. “So you etch copper out from underneath and have graphene directly supported by a porous polymer – which is basically a membrane.”
The researchers performed diffusion tests with the graphene membranes, flowing a solution of water, salts, and other molecules across each membrane. They found that overall, the membranes were able to withstand the flow while filtering out molecules. Their performance was comparable to graphene membranes made using conventional, small-batch approaches.
The team also ran the process at different speeds, with different ratios of methane and hydrogen gas, and characterized the quality of the resulting graphene after each run. This gives a great degree of flexibility in terms of tuning the graphene for electronic or membrane applications, say the researchers.
“In the end-to-end process, we would need to integrate more operations into the manufacturing line,” Hart says. “For now, we’ve demonstrated that this process can be scaled up, and we hope this increases confidence and interest in graphene-based membrane technologies, and provides a pathway to commercialization.”