Irma Kuljanishvili describes nanoscale fabrication research of semiconductor devices.
Focus on Dr. Irma Kujanishvili interviewed by Stephen Ogg
DR. IRMA KULJANISHVILI
Dr. Irma Kuljanishvili, is passionate about integrating nanoscience into technology. In a recent telephone interview, Dr. Kuljanishvili translated her research interests, their importance and their potential application from the language of pure physics into terms that scientists from other disciplines could more easily understand. As an applied material physicist, she is interested in new ways to create and modify materials that are important for technological innovation. Specifically she uses her expertise in graphenes and carbon nanotubes to investigate their potential incorporation into future devices that will find applications in widespread fields, including semiconductor manufacturing and quality control as well as other sensing and testing applications. Dr. Kuljanishvili uses a variety of standard scanning probe techniques in her laboratory, but also is pushing the technology delving into development of scanning probe techniques themselves. As these techniques to interrogate nanodevices are developed and mature in research labs such as Dr. Kuljanishvili’s, they will transition to production facilities where they will be integrated into quality control of new generations of nano-scale devices rolling off of assembly lines.
Can you describe your research problem?
Our group specializes in carbon nanostructures and devices. Although we are an applied physics group, we also make our own materials. The two main materials we’re interested in are carbon nanotubes and graphene. We also work with different types of thin films, but that’s not our primary interest. We synthesize our own materials and we have a unique way of handling catalyst particles on substrates to grow our materials. Most techniques that we use in the lab in one way or another use scanning probe microscopy. For example, if we would like to create and grow carbon nanotubes, catalytically grown on the substrate of silicon for example, we force pattern the catalyst particles. We derive them from solution. We generate a solution from say, iron or nickel containing salts. We dissolve them in solvent and then to pattern we use a technique that is based on scanning probe microscopy, very reminiscent of DPN (dip pen nanolithography) to pattern the catalyst. We do this with an AFM, but we also have a stand alone, home built system that nevertheless uses the scanning probe microscopy technique to create clusters of catalyst on the substrate for the growth.
Why can’t you use normal light based lithography techniques to pattern the catalyst?
Resolution of traditional, conventional lithography, depending on which type of lithography you use, photolithography or helium lithography, is a multistep process that requires facilities that need to be used on a daily basis to perform. Resolution wise, with helium lithography you will be able to get resolution down to the nanometer scale, or sub-nanometer scale easily. But it’s mainly not about resolution, mainly it’s about having one step, directly using AFM probes to write, without using any sort of mask, or any sort of process in between. You write your catalyst where you want it and you grow your materials only at that location. So we call it predefined, preferentially grown at a specific location. This is something that is still very desirable for device applications. You can create arrays of devices. You would clearly like to have your carbon nanotube grow only where you want it and not have catalyst at other places where you don’t need it. Plus it eliminates the multiple step process required in convention nanofabrication. The process that is usually used, if you really want to create islands of catalysts and grow from there and the one we use is using AFM probes.
What volume of droplets can you deposit onto your substrate using scanning probe techniques?
It is usually in femtoliters (10-15 liters). Everything is femto, volumes are femtoliters, measurements are femtofarads, charge is femtocoulombs.
Can you manipulate single catalyst molecules?
No Not at this point. At this point when we deposit catalyst in molecular solution, it is still in solution. It dries out and we reduce it. We still get, at the end, a sample of catalyst particles. Statistically, the smaller the cluster, the fewer individual particles that will be in the cluster. We keep them as uniform as we can. Statistically at this point we are growing individual carbon nanotubes from the cluster of catalyst containing, perhaps, tens of molecules. Then we use chemical vapor deposition technique, which is pretty high temperature thermal furnace that we grow carbon nanotubes in. We have been playing around with devices that are already pre-fabricated, like pre-fabricated transistors. We collaborate with researchers at Argonne National Lab, they provide us with devices and then we put catalyst on the device in specific pre-determined locations and then we grow our nanotubes. The downside of this is that the material and leads that are still metal degrade a little in the furnace at 900°C. You have to be clever, either using a lower temperature so your metallic components don’t degrade, or use some other metal that still preserver their integrity even after high temperature processes. Most devices do not survive CVD conditions. That’s another area of research. Eventually, perhaps, we will go to a different regime or schematic, or use different recipes for significantly lower growth temperatures, which could be ideal.
What types of devices do you hope to build/construct with the materials that you are growing?
We are interested in different types of devices. Basically those are field effect transistors. We would also like to create 3 dimensional structures. Imagine that we build a device – a planar field effect transistor from individual carbon nanotubes, or individual piece of graphene, later we can, depending on the size of the device, with scanning probe lithography, or patterning, we can deposit molecules, DNA, proteins, or other particles, on top of your ready device and modify it this way. The important thing when we work with graphene or carbon materials is not to modify the structure through covalent bonding. What we would like to do to modify, to some extent, with chemisorption or physisorption so it could be reverse processed, or it could be a process that doesn’t permanently destroy the crystalline structure of graphene. That’s what we also do with the surface of graphene and we’ve demonstrated patterning of magnetic particles on the surface of graphene with our AFM based direct writing techniques.
Why is your research important?
We’re still physicists. We really would like to learn ways we can create and modify materials that we make. Some materials, as grown, are perfect and can be used for specific applications. Some materials have extremely unique properties, like graphene. One property that graphene doesn’t have is inherited band gap in its electronic structure. But you can create band gap opening if you modify graphene without really destroying the structure, without creating defects and use this as a sensor. We would like to use modified graphene with organized arrays of patterned structures, whether it’s kinetic/magnetic/233 or else to preserve its integrity and also is modified to be able to be used as a semiconductor.
What techniques do you use in your research?
We use techniques that relate to nanoscience and nanotechnology. We are now more and more involved in chemistry and become a cross-disciplinary group, so we’ll see. But our goal is still to work with devices and try to understand how modified devices behave, how reliable they remain after process of testing. I think that’s how ultimately anything we study or discover becomes an application. If it’s really something that reliably performs after many cycles of treatments, whether it’s cooling it down and measuring conductivity, or measuring other characteristics, or doing spectroscopy on the device, we would like to know how well the device survives.
How is AFM used in your research?
Well, we need nanoscale observation to study nanoscale devices. We use AFM, and of course non-contact is better, to observe the nanoscale materials we are making. Non-contact AFM allows us to “see” our materials without perturbing them. Ultimately we need to find out the behavior of our devices without destroying them, so we need techniques that are non destructive. When your materials are so small, the chance that the observation will change the device is something that we have to be careful of. Also as you shrink your sizes, quantum behavior becomes more prominent and so we utilize other scanning probe techniques to measure physical properties of our materials. We cannot avoid contact when use electrostatic force microscopy to measure the electrical responses of our devices. I’ve described our DPN technique to you already that we use for patterning catalyst to grow carbon nanotubes, but we can perform localized annealing or heat treatment on the nanoscale. We also use magnetic force microscopy to measure magnetic properties graphene and carbon nanotubes, which are not magnetic, but after we engineer graphene with a high density magnetic field.
What can typical consumers hope to see in the future that would be derived from something you’re studying now?
We really hope, with the studies that we do, we get something interesting discovered. For example, if we can make a reliable sensor, or a reliable transistor out of graphene that has been modified – for a consumer it probably means faster components of electronic devices, or nanoscale electronic devices, smaller nano-electro mechanical devices. Testing and sensing applications can be performed on the nanoscale. Perhaps it could also mean a new type of diagnostic probes that we can also manufacture on a small scale. Imagine that you need to test or investigate a small sample or valuable specimen that you have very little of. You can test it in a traditional way, but you need big volumes of it. With AFM techniques you can deposit tiny portions (femtoliters) of something on a glass slide and study it this way. I’m thinking of specimens, especially biological, DNA based or protein based, that you could do this with. This means that a lot of tests could be performed with a much smaller amount of specimen that you would currently need to provide. This could also be applicable to devices for all other fields, for example semiconductors, but biology is a natural fit, simply because things that you can do in solution are easy. You can manipulate this and have really tiny volumes, in femtoliters.
I have a colleague in biomedical engineering department and right now we are trying to develop a project for a mutual student who will work with carbon materials and cells. Using scanning probe techniques, we won’t be able to manipulate micron scale objects like cells, but we can try to get a nanoscale response from a cell that is attached to a device. We can study a cell that sits on a patch of graphene or a carbon nanotube and record its electrical response or its capacitance response. Actually, if it is already sitting on a device, we can also record conductivity while simultaneously stimulating the cells by releasing molecules from within the carbon nanotubes at specific times. My colleague is very excited at the prospect of studying responses of individual cells.
But you don’t have to apply our research to biology. For example, semiconductor manufacturing is currently a bottom up approach and we hope to achieve similar thing at the nanoscale. But really, we want to integrate nanoscience techniques into technology.As we develop methods of fabrication on the nanoscale, these devices will have applications everywhere. Anything with a sensor, or integrated circuits can be miniaturized and therefore require less sample or potentially become much less costly to produce.