This week I spoke with Giovanni Oakes, a PhD student here at the Cavendish Laboratories. Giovanni’s work primarily focuses on CMOS quantum devices as well as drinking an abundance of coffee! Hello Giovanni!
Could you tell me a bit about yourself and how you found yourself to be at Cambridge?
Hey Tara! For those of you who don’t know me, I’m Giovanni, I'm half Italian, half British and I was brought up in Italy in the land of prosecco until I was 15. That's how I famously introduced myself when I joined the NanoDTC! I then moved to the UK, where I did my undergraduate degree at Imperial College London in Material Science. Like most NanoDTC students here, at the end of my degree, I wasn't sure what I wanted to do, but I was certain I wanted to keep doing research. I had various research experiences, some in Cambridge, in the Materials Department and I knew I wanted to do something similar.
What research experiences had you completed before coming to Cambridge?
I previously worked in Jason Robinson's group, focusing on superconducting materials. I was looking at spin valves for spintronics applications. And whilst I was there, I was shadowing Angelo Di Bernardo, who was a NanoDTC student, and he pointed out the DTC to me at the time when I was sort of figuring out what I wanted to do. And yeah, long story short, I applied and got in!
I also did an internship at IMAC in Belgium. It was a four-month internship that had to do with my integrated master’s. We had to do an internship and I got to work at IMAC in a €200 million clean room, which I thought was pretty cool.
IMAC is a weird mix between research and company, and many big semiconductor companies will ask IMAC to do research for them because they have a clean room set up for research. A company sponsored the research I was working on, and to this day, I still don't know which one it was. And because I didn't sign an NDA with that company, but just with IMAC, I was not allowed to know what the project was for and who was in the project! It was strange!
So you were just doing all these measurements that you had no idea?
Yeah. So I was basically just told that they had figured out how to epitaxially grow Indium Arsenide on silicon. And the next step was to figure out what dielectric you could deposit and at what temperatures it could tolerate, because often in silicon, you anneal at very high temperatures to improve the dielectric. But if you heat up Indium Arsenide, you'll form arsenic gas, which is not great! So, I was basically just in a cleanroom, depositing random dielectrics and annealing and seeing what happens and then reporting back how badly it degraded. There was a lot of fabricating, just simple capacitors and then doing CV
measurements to figure out the number of defects on the surface. While I was there, I also got to learn to use an AFM and XPS.
So, was it an enjoyable experience overall?
I enjoyed it overall, but I learned I didn't want to do a fabrication-based PhD because it's a lot of repetitive work and I was more interested in the science rather than optimise a recipe.
That’s good to know! Can you tell me a bit about your first year and how you decided what PhD project you wanted to do?
Well before coming here, I did something very different for my master's project at Imperial, which was DFT on perovskites, because I wanted to see the computational side of things. I also had various lecture courses on batteries and renewables. And I thought, oh, I'll do energy materials. It sounds cool and useful. It was good to have a background in DFT to know what it is capable of and its problems, limitations and whatnot. So, when I started, I thought I was going to do something with energy materials. I told Karishma, our director, about my interests and the first project I had was to do with thermoelectrics with Chris Ford's group. And then the second project was with Jenny Zang on bio-photovoltaics, and I had never done anything with bio before! I think Karishma likes to put people really out of their comfort zone, so I thought that was fun.
How challenging did you find the bio? Was it a different experience from previous labs?
Honestly, I thought it was such a messy environment. I found it so hard to reproduce data. Don't get me wrong, I think what biochemists and biologists are doing is very interesting and I think there's a lot of untapped potential in bio which is only recently being discovered. It is a booming field in science at the moment, but it wasn't for me. The group I was working with were lovely and incredibly helpful but after the project, I knew bio was not my thing!
How did you go about choosing your PhD project, because I know it's very different to batteries and energy science!
At the end of the MRes year here at the NanoDTC, I still don't know what I wanted to do, and I feel like my whole cohort didn’t have clear ideas about what we wanted to do. But one really nice thing about the NanoDTC is that they provide you with a list of potential projects and you can choose which projects you're interested in. When the list came out, I went to talk to supervisors about various projects that sounded interesting to me at the time. There was one project in particular that caught my eye, which focused on quantum devices and machine learning. I had friends from my undergraduate who were working in quantum information, so was sort of aware of that field and it sounded cool. I had no experience in machine learning, but I liked the fact that there was some sort of theory behind the simulations, which I had some experience in from my background with DFT. They were both new fields to me but I decided to choose that project in the end because I thought it sounded interesting.
So, you were saying there that it's like two new fields you're jumping into for your PhD project. That sounds terrifying, especially because some people going into their PhD project, already know a lot about the field. Were you nervous? Did you approach your PhD differently at the beginning?
Not really! I just knew I had to read up a lot. And I'll admit it took time to feel confident in the field. I feel I've learned the parts of the field that I've focused on, but there are still so many more things to learn, which is nice. I enjoy that aspect of constantly learning new things.
Okay, cool! So, you mentioned your project is on ML and quantum devices. What is your project particularly focused on?
My project focuses on understanding and designing a type of quantum device. The main idea is that we are very good at fabricating and scaling up silicon-based transistors. And the question is, can we operate them in a different way or in different conditions that can be used for quantum computing? In particular, the idea is to cool down effectively just normal transistors to very cold temperatures around 11 mK.
Given the right conditions, you can trap single electrons under the gate of a transistor, thus forming a quantum dot. As you apply a voltage, you can load one electron at a time. Basically, you can control where your electrons are! It's no longer Ohm's law, but rather you can model it as a semi-classical coulomb blockade where there's a certain energy penalty to add a single electron to this dot. This is because the added electron feels coulomb repulsion from the other electrons already present. By manipulating these electrons, we can effectively use them as qubits, where the spin of an electron (up or down) can be mapped to a bit of information (0 or 1).
You mentioned a ‘qubit’? For the readers who don’t have a background in quantum mechanics, could you explain what a qubit is?
A qubit is short for a quantum bit, which is the equivalent of the classical bit in a classical computer, and it stores information as a zero or a one. The added advantage is that you can exploit quantum phenomena, such as entanglement and superposition, to create operations that you can't have with a classical computer.
This sounds very powerful... Then why isn't everyone doing this? Why don't I have a quantum computer?
At the moment silicon quantum dots are still in their infancy stage, in the sense that the largest silicon/silicon-germanium quantum dot array consists of just six quantum dots. That has been published quite recently. And in my group, we're still trying to get one qubit to work!
So, you're making all these qubits on silicon chips. Are there other forms of making qubits?
Yes. Anything that is a quantum two-level system can potentially be used for quantum computing. Currently, one of the most popular platforms is superconducting qubits using Josephson junctions. That's what is currently being pursued by Google and IBM, for instance, which is what you often see in the big headlines. But then there are other platforms like nitrogen vacancies in diamond, silicon photonics, ion traps, to just mention a few.
Why focus on silicon qubit devices, as opposed to other types?
We don’t know which technology will end up becoming commercially viable, so we need to experiment with all these different platforms. The appeal of quantum dots is to leverage the silicon industry. However, there are different parameters that you want to optimize between a classical and a quantum dot transistor. You are susceptible to different types of noise. And there is a learning curve that has to be overcome in terms of processing, which is also why there's a bit of a struggle to create these devices. But the main idea is you'd be able to scale up and it doesn't take as much physical space on the chip, so you can potentially fit billions on a single device.
Quantum computing is such a broad field in itself. What does your PhD project focus on specifically then?
I started off with more of a computational project, where I was building simple models to simulate how each gate affects its neighbours and creating some synthetic data. I then developed a machine learning protocol that extracts the parameters required to control each quantum dot independently. But as my PhD progressed, I began measuring devices and the project became more experimental. The group I was working in are very good at readout, i.e. measuring the state of the quantum dot.
During my PhD, I found a transition that had a singlet-triplet anti-crossing and was able to do single-shot readout of the spin. I managed to measure the highest fidelity single shot readout in the shortest time using a single electron box as a charge sensor, with 99.2% fidelity in under six microseconds.
That's pretty amazing.
Yes, that was fun and completely not what the PhD was supposed to be about, but that's what makes a PhD fun.
That's one thing I'm noticing throughout interviewing everyone is that the PhD starts off with this proposal, but they always managed to engineer the project themselves and kind of deviate away from that. By the end, their project is nearly something completely different. Have you found that as well?
Yes, these results were definitely unexpected. The initial idea was to experimentally test the model I developed to automatically tune a two-by-two array of quantum dots. But the devices we had access to in the lab did not work, and I spent so much time on these problems, so we had to deviate from the project and reshape it in a way.
How did you kind of deal with the frustration of working so hard on one thing and not working out? Because PhDs, they're just kind of full of up and downs. How do you deal with the downs?
I think you just have to persist and know that you are doing research, meaning that you're working with a technology that isn't fully developed. Things are going to break, things are not going to work. But at some point, you have to pursue something until you feel that it might be that it is just a dead end, which was the case for me. But it's worth knowing if that's the case or not. And then you can move on. And I think you have to have some sort of a plan B.
Throughout your PhD, you've worked with several industries, including Hitachi and Quantum Motion. Could you tell me a bit about how the industry side has incorporated into your PhD?
To start with, I was affiliated with Hitachi Cambridge Laboratories. They're the ones who have provided throughout the PhD samples and the equipment for me to measure and effectively have been my day-to-day supervisors. During the COVID period, my industrial supervisor, Fernando Gonzalez Zalba, decided to move jobs to Quantum Motion which is a startup based in London. As well as this, my main academic supervisor, Charles Smith, decided to step up and become the lab manager of Hitachi Cambridge Laboratories. And I was given the choice of whether to work with Fernando or to have a new supervisor in Hitachi, who they still had to hire. I decided to continue with Fernando, but I still have kept close ties with Hitachi. It's very interesting also having the two different perspectives from both companies. One is a very well-established global organisation. There's a lot of bureaucracy within the company and it's interesting also seeing how the company has a lot more money to invest than a startup. But it's also hard for them to convince people higher up the ladder that they need to invest. And it's a lot of work and paperwork in order to, for instance, buy a new cryostat and this sort of stuff. On the other side, you have a startup where it's a lot easier to get equipment and resources. One is definitely more flexible than the other, but the other has potentially a lot more resources. It's just whether you can get people higher up the ladder convinced that it's necessary.
Okay, I never really thought about it. I like that!
In fact, I had the opportunity last month to go to Japan and visit the Hitachi labs in Tokyo, and that was very interesting. We went for a week of meetings, basically, in the research centre in Tokyo, and yeah, I mean, they definitely have resources. They have things that I feel a startup doesn't have access to. For instance, they have their own clean rooms which are expensive to maintain! It was very interesting because they definitely have a very strong manufacturing and CMOS engineering side of the company, and they understand very well what can and cannot be physically fabricated. But then there's no one there, really, that has much of a quantum computing background that understands what are the key parameters that need to be optimized. And the idea is that the research group here in Cambridge are those physicists who have a quantum background and are supposed to report back to Japan, which isn’t easy with an eight-hour time difference and a language barrier.
Sounds like an exciting trip! So now starting to finish up, and you're writing up your thesis, I think you're saying you have two months. How is writing the thesis getting on?
Yeah, I think there are some chapters that have been easier than others, especially the ones that you have at least some sort of a paper or paper draft written for. It's been nice to have time to go through certain things and try to explain them and make sure that I myself understand them properly. I’ve been trying to run simple simulations to make nice figures, which I think has been also quite beneficial for me and in a way might be the most useful part of a thesis.
And finally, do you have any idea about what you want to do after your PhD?
I'm definitely going to stay in this field and see where it goes, and work with a research based industry. I don't think academia is really for me overall. Postdoc positions tend to be one to three years max and you're just constantly applying for the next postdoc and having to move around the world. And I think that's just a lot of stress. I would much rather work in a company which I think is more secure!
Thanks Giovanni for the chat and all your advice! And good luck with the write up!
Special thanks to our Editor Larry Brazel for his work and advice.