We’re always hearing about how graphene is going to revolutionize this or that. Has progress in research and commercialization been as quick as you’d anticipated?

I think so, because if you look at any major technology based on hardware, as opposed to software, it has always taken between 20 and 40 years to go from the initial discovery to mass applications. The first transistor was in 1947, but it took more than 10 years for people to realize that they had to use silicon instead of germanium. Or if you take carbon fibre, the first time anyone put carbon fibre in a car was in the earl 1980s, but the first commercial fully carbon-fibre car only happened in 2013. Any new “node” technology requires several years to develop and massive investments of (at least) $10bn, and that is just for one node.
So yes, it’s completely normal to expect a 20- to 40-year time frame. What is not normal is that the media has built up an expectation around graphene that is clearly not possible. It’s similar to when the government announces that there will be a high-speed rail link between two towns, and after two months people ask, “Where’s the train?” It clearly doesn’t work like that. From my point of view, we are perfectly on track over a 20- to 40-year roadmap for graphene applications.

What will those first applications be?

The first applications are already here, and they are related to inks, coatings and polymer composites. There are skis, tennis rackets and cars with graphene in them that you can buy, and masks with graphene filters for working in the lab. I mean, there’s lots of things you can buy already that are made with graphene. Even underwear.

Graphene underwear?

Yes, you can buy graphene underwear – there’s a company in China producing it. There are many, many products out there. If you want a motorcycle helmet with graphene, you can already buy that. You can buy shoes that have graphene inside. These products are not particularly hi-tech and they mostly rely on graphene’s thermal properties, maybe antibacterial properties, and mechanical properties. These are the low-hanging fruits and we now have a plethora of start-up companies in the UK and also in Europe making graphene inks, composites, paints and powders.

But then you start to think about creating consumer electronics – optoelectronics, photoelectronics – with graphene, and this will require massive investment. We’re talking about 10 or 20 years before we can really implement these. Yet the more it goes forward, the more I start to believe that the technology will succeed. I cannot tell you if it’s going to be a million-, billion- or trillion-dollar market, but for a million-dollar market we are already there – it already exists.

These great investments that are needed over the next couple of decades – what do they need to address?

The investment required is of a similar order of magnitude to what you need nowadays to develop a new silicon-based device. We are talking about hundreds of millions to develop a technology that is ready for mass production. But you can also have much lower investment in the form of pump-priming to help companies test proof-of-principle devices.

So is the money going into designing and building new fabrication facilities?

No, the idea at the moment is to integrate graphene with existing silicon CMOS fabs. Of course, there are new tools being developed, but essentially these are chemical vapour deposition tools, so they are very similar to the tools that you already have in the fab. What you really need to develop is the process – the temperature, the chemicals and also things like the fact that at the moment graphene is often grown on copper foil, which is forbidden in silicon fabs because of contamination. You can use iron chloride to remove the copper, but then you get iron contamination. So one direction is to try to grow graphene directly onto silicon – this is very complex, but probably feasible.

We are entering a phase where a lot of things that are not so important when you write a paper for a scientific publication are becoming crucial. If you want to develop this technology, you need to optimize the material to meet the key performance indicators, such as electron mobility and scattering time, that will really make graphene much better than the current technology on the market – and you need to do it in a way that will work on a mass scale. In the lab, we have already shown that graphene is much better than the current technology. But clearly, a laboratory device doesn’t face all the problems that you get when you need to scale up the device and integrate it with other systems.

Where do other 2D materials fit into this picture?

I would actually call them layered materials, because while graphene is certainly a monolayer of atoms, for something like molybdenum disulphide (MoS₂) the “one layer” is actually three layers: one of molybdenum and two of sulphur. So it’s a bit of a stretch calling them 2D, even though, clearly, they are more 2D than 3D. There are about 2000 of these materials that we know of and very few – probably 10 or 15, although I don’t know the exact number – have been studied so far. Of course, people have studied many of them in bulk form, but you have to exfoliate them to get the monolayer, and when you do that, the material’s properties change. The other thing is that even with graphene, when you stack a layer of graphene on top of another layer, you can go from a metal to a semiconductor just by changing the angle between the layers. The same thing happens with all the other materials, and when you put them together – say graphene, boron nitride (BN) and MoS₂ – that creates yet another spatial option. And then you can twist them, and so on.

So, in terms of research options, we have plenty of materials to study for the next 100 years. From the applications point of view, though, clearly graphene has come much further than the others because work on it started earlier. So when we speak of applications, graphene is going to be the first. After that, other materials that are now reaching maturity include BN (as a nice substrate or a way of encapsulating graphene), MoS₂, and tungsten diselenide or tungsten disulphide (WS₂). These materials are getting closer to the application stage and if you are interested in light-emitting diodes or even transistors, they are better than graphene because they have a band gap and graphene does not.

These other 2000 materials – you say they haven’t been studied yet, not because they aren’t promising, but just because we haven’t got around to it?

Just because we don’t have 200 million PhD students, yes. It should also be said that until two years ago, nobody bothered to classify these layered materials. Now, several theoretical groups have used data-mining techniques to search the crystallographic databases for layered materials, and they have established that there are between 1500 and 2500 of them – it’s not even clear what the precise number is. There are also at least 80 other materials that are three-dimensional but can be made layered; you start from a covalently bonded material that you would not think could be exfoliated, but then you do some tricks and it becomes possible.

So, really, this idea of graphene has uncovered a completely new field of research. Almost every week there’s something new coming up. But from my point of view, at this moment I would rather take graphene and WS₂ and BN and so on – these few materials that are well known – and really push them forward technologically. I think new groups may have a bigger incentive to go and explore random materials because I am pretty sure that at some point somebody will find a material we don’t know about that has much better properties than graphene and the others. But we can’t keep looking for the new ones if we are interested in applications.