Dr. Kiran Chikkadi is currently a postdoc at the Chair of Micro and Nanosystems Group at ETH, Zurich. He completed his Bachelors in Electronics and Communication from National Institute of Technology, Surathkal, Karnataka, India. He joined the Masters program in Micro and Nanosystems at ETH Zurich and further enrolled for PhD in the same group. His PhD thesis mainly focused on Nanotube gas sensors and process control monitors for batch fabrication of Nanotube-based sensors.

Can you give me a brief introduction about your background and work so far?

For my PhD I mainly worked on developing Carbon Nanotube-based (CNT) gas sensors. Just to highlight the significance of these sensors when compared to other sensors in the market, normal sensors consume about 10-100mW of power and MEMS in the order of milli watts. Additionally, both operate at higher temperatures. Carbon nanotubes consume nano watts of energy and operate at room temperature. So right there you have a difference in the order of magnitude of about six-to eight with the advantage of operating at room temperatures. These sensors could be used for breath monitoring, pollution monitoring, explosives detection, etc. The sensors I worked on were mainly used to detect NO2.

What is the simplest part about these nanotubes?

Well, we know that they exist (laughs) considering their very small size. But apart from that, they are semi-conductors with well-established physical properties. For me, the fabrication process is the simplest part.

What are main problems associated with developing these sensors?

My work mainly involves development of the device and generating data with them. There are several problems that we face. First, the tubes are synthesized on a silicon/silicon oxide substrate and there is some inherent imperfections associated with the silicon oxide. This introduces noise. Secondly, there is a large variation in the length of the tubes. There are infinite possible varieties of tubes that can be grown. This mainly arises due to the possible ways in which graphene can be rolled up giving rise to different structures. Lastly, there is density variation in the nanotubes grown, i.e. the tubes grown are a mixture of metallic and semiconducting devices. Since, we are interested only in the semiconducting structures, sorting them is not easy. In short, there is a random distribution of tubes and picking the right tubes is extremely difficult.

So where do you face complexity in your field and how do you deal with it?

For me, complexity arises from simple phenomena. In my field it is basically noise. To be precise, it arises from “pink” or 1/f noise. To improve the performance of CNT devices, the knowledge of their noise mechanism is crucial.

The noise mainly originates from the charge traps surrounding CNTs. The electrons sometimes occupy these traps in the substrate and can either remain there or can jump back. Their constant switching can be observed as fluctuations in the current being measured. These measurements show that the low frequency electronic noise is dominated by 1/f noise or pink noise. Now, this pink noise and complexity are intricately connected and this noise is not limited to CNTs. It is ubiquitous. This is one of the oldest problems in almost all fields still lacking accepted explanation. This pink noise is an intermediate between white noise and Brownian motion i.e. state history independence. Despite observing this noise in almost every field there has not been a generic mathematical explanation to describe this noise. There are many scientists in the field of Complexity trying to model the same. Thus, the unique yet pervasive nature of this noise makes it difficult to eliminate it completely. But there are some ways we could reduce it. For example, we found out that we could strongly reduce the noise contributions from the traps by suspending the CNTs. This is simply because the CNTs are not in contact with the substrate any more. Of course there are other sources for pink noise but this is one of the ways we deal with it.