Panel Discussion on Carbon Nanotubes
Carbon Nanotube Session, MRS Memorial

Photo of front row of NT13 group photo
Photo of front row of NT13 group photo

Esko Kauppinen: We have enough time, now, to have some discussion on different topics and hear the opinions of some of the world's leading scientists in carbon nanotubes.  We should always be thinking about the future, of course, and materials creation always contains three steps: synthesis, characterization and application.

So my first question for all of you is: what do you think are the challenges for the future? It could be the synthesis or characterization or application in the coming five to ten years. Let’s focus mainly on nanotubes, but you could also comment on other 2D materials.

Photo credit: Shoshi Dresselhaus-Cooper
Photo credit: Shoshi Dresselhaus-Cooper

Hui-Ming Cheng: I think there are two biggest challenges for the future. The first challenge is controlled synthesis.  As Jin Zhang said, targeted synthesis may lead to predictable growth. But I think that focusing on targeted growth may be more important, if you’re thinking in terms of applications. The second challenge — the greatest challenge — is applications.  As we know, carbon nanotubes have attracted an ever-increasing amount of attention since 1991. According to Endo-sensei, the nanotube has a much longer history than many other 2D substances, but we are still not widely using carbon nanotubes in industry. That is troubling.  We need to make efforts towards the applications side of nanotubes. Right now, in China, multi-walled carbon nanotubes are being widely used in lithium-ion batteries — with several hundred tons per year. This is a great application, but graphene materials are now threatening to overtake nanotubes in lithium-ion batteries.  We need to think of more applications. Some years ago, we thought that a transparent conducting film might be a very good application. But right now, it seems that the cost of creating this material would be a big issue in terms of mass production — and we’d need to get that down. Overall, I think we all should consider these two challenges whenever we do nanotube research.  These two challenges are very… challenging.

Esko Kauppinen: Okay, so we have challenging challenges. Maruyama-sensei, please, your comments.

Millie and Rick Smalley at the 1997 ONR Workshop on Carbon Nanotubes
Millie and Rick Smalley at the 1997 ONR Workshop on Carbon Nanotubes

Shigeo Maruyama: I believe that all three — growth, characterization and application — are very important. When we come up with very nice applications, we suddenly realize that the nanotubes are not grown to support that application, so we all come back to the growth issue and, suddenly, growth becomes very important. So they’re all interrelated.

Of course, we have been working for maybe 25 years on the growth issue, and we probably still need some major breakthrough in that area. We are not completely confident in the mechanism. My point is, growth is very important.

But characterization is also important. Some student just began to talk, outside, saying, “2G prime?  What are they talking about?” I’ve found this a very nice way to check how many papers my students have read in the field. But the problem is that actually, this topic is very complicated and convoluted — and perhaps it would be to all our benefits if someone could write this all down in a very easy-to-understand way.  It’s not easy. In fact, it’s made harder by the fact that many in the field have read such technical papers for so long that we’ve grown used to it. After all, we have 25 years of papers like this! But I really do believe we need a simple, uncomplicated summary of characterization so that we can allow people in other areas to be able to read it, understand it, and use it in the next stage of research.  That would be a very important development.

Overall, I believe there are many interesting applications and collaborations with industry that are being set up right now in China and Japan, but we still have a lot of things we need to do to make nanotechnology more widespread, industrially.

Esko Kauppinen: Marcos?

Marcos Pimenta: Some 20 years ago, when we first started to work with carbon nanotubes, we first considered a sheet of graphene, then roll up the graphene into a tube and introduce new phenomena — basically, the quantum confinement. But at that time, we spoke of graphene as a kind of toy model for us, because it didn't exist in nature.

Then nanotube research exploded and, a few years later, it became possible to isolate one single sheet of graphene.  And after that, nanotubes were not considered such a hot topic, anymore — and graphene became a very hot topic.

Millie took apart her graphite lattice structure models to create graphene.  Photo by Shoshi Dresselhaus-Cooper
Millie took apart her graphite lattice structure models to create graphene. Photo by Shoshi Dresselhaus-Cooper

But it was strange to think about, because — why did nanotubes appear before graphene?  We know that the nanotube is much richer in physics and applications than graphene, since graphene is 2D but a nanotube is only 1D and has quantum confinement. It can also be used in interesting ways, as its electronic structure depends on the chirality. So why are graphene and 2D materials the current ‘hot topic’, instead of nanotubes?

I think the main reason for this is simple: we don't yet know how to produce good samples of nanotubes that have only one single chirality.  I believe that this is the main thing preventing nanotube science and technology from going further.

So the challenge for the future really is to produce samples with just one single chirality. Then, if you can grow nanotubes of a single particular chirality, all future experiments will become much simpler, and industrial applications will become cheaper. And at that point, I believe, carbon nanotubes will be a hot topic once more.

Okay. So I think the challenges are very clear. The obvious is very clear. For my opinion, it’s the same, I think.  For the first thing, synthesis. We always say that synthesis determines the future. If you want to use these kind of materials, if you cannot get these kinds of materials, you’re seeing no applications.

Millie demonstrates graphene becoming a nanotube. Courtesy of the Dresselhaus Family
Millie demonstrates graphene becoming a nanotube. Courtesy of the Dresselhaus Family

Jin Zhang: My opinion is very similar to everyone else’s.  I think the challenges in this field are very clear.  Let’s talk about the first challenge: synthesis. We always say that synthesis of a material determines the future of that material. This is common sense: if you want to use a material, you have to know how to produce that material — otherwise, any applications you come up with will just be pipe dreams.  Therefore, if we want to create applications for carbon nanotubes, we first have to develop some way to control the growth of nanotubes. Also, we have to be able to grow high quality nanotubes in bulk.

Characterization, however, is also very difficult.  Extremely difficult! Both for individual tubes and for an array of tubes, we run into repeated problems with characterization. It is true that you could use TEM to look at that a single nanotube — that’s correct, to a certain extent — but at the moment, even this is difficult to accomplish. Therefore, I believe we should develop some sort of scan that can more easily characterize nanotubes.

Photo credit: Shoshi Dresselhaus-Cooper
Photo credit: Shoshi Dresselhaus-Cooper

The third challenge is, of course, in applications.  I think that we have to look for some kind of killer application for carbon nanotubes. We know that carbon materials can be very useful in applications — just think of graphene, and how many applications it has, even now! I’m not saying that carbon nanotubes will, necessarily, have applications in the exact same areas as graphene, but the raw material shows potential for big applications.  If you think about silicon and all the applications that have been produced for that material, you can see that the potential of raw materials is an important factor. I believe that, in the future, we will be seeing carbon nanotubes and graphene used as widely as silicon is, now.

So, in short, we must address all three: synthesis, characterization, and applications.  That’s it.

Esko Kauppinen: Okay, so I’d like to add my own comments to this discussion. It’s true that I didn't have a talk, but I've been working on carbon nanotubes in each of these areas — synthesis, characterization, and application.

Millie reads science books in Japan. Photo courtesy of the Dresselhaus Family
Millie reads science books in Japan. Photo courtesy of the Dresselhaus Family

I think it's important to “make the loop”, so to speak.  In other words, we must focus on all these areas, because they are all crucial steps in creating a material that is commercially viable. First, we make the material, but once we make the material, we need to characterize it so that we can use it in applications and industry. But of course, it’s easier to get funding to make the material if industry can see the applications. So the process really is like a circle.

Recall that even if we make the (1n, m) tubes, it doesn't guarantee we have a good application. It opens the possibility, or maybe does a little better than that, but we still need to complete the circle.  This is our challenge with nanotubes.

Look at graphene for a moment.  It’s not easy to make a really good quality sheet of graphene.  We rarely have perfectly clean and ballistically conductive graphene. Well, the papers are all written assuming we have that, but in our labs, we rarely do. That’s a big difference.

Of course, it’s easy to say synthesis is the key, but that may be jumping the gun.  We must approach the entire circle in an iterative way. Perhaps now, it’s more important to think about the characterization. Earlier, it has been more like Millie’s legacy of photophysics to understand the features that we have. But then, if we use these methods to look at the sample and characterize it, the characterization is not simple — it’s not like we feed a nanotube into a machine that analyzes it and then goes, “Ding!  It’s an armchair nanotube!” Instead, we get distributions, which are complicated to read and understand, and that’s a problem. We need to make this all simpler, faster, and cheaper.

As we know, samples are always distributed. It’s the same with graphene and 2D materials; we have distributions and always will.  It’s a reality we cannot avert. In basic science, this is a nuisance because we want to have just one simple solution, but in the practical world, there isn’t just one simple solution — it’s always going to be more like a distribution.

I’ve been working with transparent thin films and there are many issues.  (n,m) is one, of course, but there are also issues like purity, orientation, length, whether or not the nanotubes bundle, etc.  It’s the same with the graphene and 2D materials; we’re always scrambling to understand the actual structure we have made.

I agree that a killer application would be needed, but I will add in a word of caution: it’s very difficult to replace existing materials. And when we’re coming up with new applications, that’s what we mainly think of — ways to replace existing materials with carbon.  We think, “Okay, we have silicon everywhere, but we want to replace silicon (in the CMOS).” Or we think, “We have ITO everywhere, and we want to replace ITO with transparent conductors.” Even though our new materials might have slightly better properties and make an improvement to the device — in the commercial world, the companies have to figure out if the cost of changing to another material (changing their machinery, processes, etc.) is justified by the small increase in better devices. And that’s a problem, because it’s practically impossible for any new material (at the beginning, at least) to be cheaper than using the existing material.  Hui-Ming Cheng was discussing this, earlier — the cost factor. ITO has been made for 20-30 years, and during that time, they have been improving it. So we’re going to have trouble replacing it at a cheaper cost.

Therefore, if we only compete with the existing applications, I predict that using carbon nanotubes in applications will be very difficult. But Maruyama-sensei taught that the nanotube network has an interesting new property because it is flexible and stretchable — and that is opening up an opportunity for new kind of applications that could create new kinds of devices that could not be built with existing materials.

I just wanted to add this to the conversation because I think this is important part of the way forward. In my opinion, carbon nanotubes, graphene, and 2D materials are facing similar problems to all other new materials.

Photo courtesy of Richiiro Saito.
Photo courtesy of Richiiro Saito.

Okay, now, moving onwards.  We have still some time, and our panelists are still eager to answer questions, so perhaps we should now take some questions from the audience.

Question Asker: I would like to ask a question in the tradition of Millie.  As everyone here knows, Millie was always very optimistic, and she liked to turn problems into an opportunities. So I’m going to ask a Millie question.  In light of the challenges that you’ve discussed and in the light of the competition from other new materials, what makes you excited about nanotubes? Why do you get up in the morning and work on nanotubes? What’s the feature or quality that nanotubes have that make you excited for the future?

Shigeo Maruyama: Originally, people were excited by just the theory of carbon nanotubes — because it’s a really amazing thing, the fact that a nanotube can be either metallic or semiconducting by just making one very small change.  And then nanotubes appeared, and that changed the field a lot. But, of course, it’s still a real challenge to grow nanotubes. Surprisingly, though, we can see them in TEM and Raman. And watching a nanotube as it grows and appears in the real world — it’s really something.  That’s why I love carbon nanotubes. They’re just so surprising and interesting. Really interesting.

Marcos Pimenta: Millie used to talk a lot about double-walled nanotubes. We know that single-walled nanotubes are very complicated; it’s very challenging to have just one single chirality. But think in terms of graphene — we have one sheet of graphene and can work out its properties, but if we have two sheets of graphene, then the properties are completely different and depend also on the twisting angle between those two layers of graphene.

In the case of double-walled nanotubes, it would be nice if we could have control of the inner and the outer tube. Millie said this often.  She put a lot of importance in double-walled nanotubes. She thought that the future of the field was to have a controlled sample of double-walled nanotubes — because if we do that, we could have new properties: the distinct structures of the inner and the outer nanotube, and also the chirality of the inner and the outer tube (metallic-metallic, semiconductor-semiconductor, etc.).  I believe there are a lot of opportunities that can be realized if we study double-walled nanotubes. But the challenge would be how to produce a well-controlled sample of double-walled nanotubes that we could study — and we must do that, first.

Esko Kauppinen: Okay. Hui-Ming?

Hui-Ming Cheng: I think this is quite easy. Why do I find carbon nanotubes interesting?  Because of its location on the periodic table! Right now, we’ve got a society dependent on silicon.  What next? If you look at the periodic table, the answer seems to be carbon. So that’s what we’re working on.

Esko Kauppinen: Okay. My answer is that even with... we cannot control it to perfection, nor can we improve a lot, but we can already make good progress in the case of real-world applications.  Also, the nanotube is a nice thing because it’s not so sensitive to whether or not it touches the surface — in contrast to 2D materials, which tend to be extremely sensitive to touching the surface. What’s more, carbon is lightweight and flexible. Now, some nanotubes are real semiconductors and some are metallic. So I believe that we really are at the beginning of understanding and controlling carbon, and that we should look for applications that cannot be done with any other materials. Therefore, I believe that there is great opportunity.  Like Millie, I am optimistic about the future.

Jin Zhang: I’m a carbon scientist. I work on carbon in all its different forms — from carbon nanotubes to graphene.  I’ve also researched back-to-back carbon materials and carbon-carbon composites. So overall, I think carbon is a fascinating element. There are so many different things that we can do with it, but a carbon nanotube is the most complicated one. So if we can control the structure, and if we can synthesize a large number of carbon nanotubes that all have one specific structure, we can probably get very nice applications. But that is hard to do, and we need to work on it.  Still, I see great promise for the future.

Question Asker: Thank you very much.

Esko Kauppinen: One very last question, please?  Yes, thank you.

Millie and Gene with Rick Smalley at the nanotube conference at Rice University, 1997
Millie and Gene with Rick Smalley at the nanotube conference at Rice University, 1997

Question Asker: It’s not exactly a question.  I would like to share a small story with you.  I was at Rice University when buckyballs were discovered — and I remember that, for many years, it was hard to do anything with them, because you couldn’t make them in bulk form. So I remember going to Rick Smalley and telling him I was doing (at the time) Fourier Transform Infrared Spectroscopy in matrix isolation.  And I said, “Why don’t we look at the FTIR spectrum of buckyballs?”

He thought about this and asked, “But how?”

And then I started thinking about it, and I realized it would take a whole year to collect enough materials to identify — because, as you know, in order to prove the structure, it’s very easy with the IR.  At the time, I thought it was going to produce a simple spectrum. But it took many years before fullerenes were ever prepared in bulk form. (Of course, once they were produced in bulk — that was when interest started to surface, again, on the applications of buckyballs.)

But, again, at the time, I remember looking at the nonlinear optical properties of C60 and C70, and while it was interesting to look at in terms of characterization, it didn’t amount to any major applications. In fact, it took many more years before we were able to use them as electron transport materials in solar cells — and that’s when they became very, very useful.

So, with the carbon nanotubes, it’s a double edged sword, because the richness is in the complexity of the science. But until we master making them in bulk form and large quantities, and until we can control their electrical and optical properties so that we can have really significant applications, it’s going to be difficult to get as excited about them as we can with single- and double-layer graphene. Well, in a way, graphene is easier to control than carbon nanotubes.

So I’d like to hear your comments on my perspective, I guess, on how things have progressed over the last 30 years.

Esko Kauppinen: Okay, because of time, let’s have just one answer to this question. Shigeo is working on both buckyballs and nanotubes. Shigeo, please?

Shigeo Maruyama:  We use buckyballs, as you mentioned, very heavily in solar cells, so although people originally thought it might not be satisfactory due to its optical properties, it has turned out to have very unique applications. It’s turned out to be a very useful material, right? So for carbon nanotubes, we are looking for many different kinds of applications, but specifically, we’re looking for a kind of application that is unique to carbon nanotubes — a device or technology that would not be possible without nanotubes.  If we can do that, it would create a new technology and a new trend, and that’s something that could change the field. I think something like this happened with C60. Its current applications wouldn’t work with any other molecule. Right?

Millie with nanotubes and buckyballs, 1994
Millie with nanotubes and buckyballs, 1994

Question Asker: But C60 is good because it’s an excellent electron transporter, so it can be used for not just solar cells, but for any device that requires exemplary electron transport. It just so happens that most of its uses have been in solar cells and (to a lesser extent) transistors.

Esko Kauppinen: I’d just like to add a comment, here: nanotubes, even without sorting, are 10 to 100 times faster than C60 as a transistor, so we don’t need perfect control over growing a nanotube to create a better transistor.

Question Asker: It depends on the application.

Esko Kauppinen: Yes, it depends on the application. That’s right. But there are many applications.  For some applications, such as replacing CMOs, we would require very narrow distribution of semiconducting tubes, so that would require a lot of control.  But in many applications, we do not need as much control to create a better device. That’s an important aspect of nanotubes: because of the metallic and semiconductor impact, you can have a wide variety of applications.

Question Asker: So the key, essentially, is finding good synthetic chemists who can make the perfect material, just in case the application demands that. And also good materials scientists!

Esko Kauppinen: And good device people, as well — and they all must work together.  It’s not easy but I think it can be done, and I encourage everyone to think positively, as Millie always did.

So it looks like we’re out of time.  Let’s thank all the speakers and all those who asked questions.  Thank you.