Transcript
Rene – Hi! Welcome to QuBites, your bite-sized pieces of quantum computing. My name is Rene from Valorem Reply and today we're going to have a deep dive into quantum hardware development. And for this, I’m very honored to have a special expert guest today, Kayleigh Cassella. Hi Kayleigh and welcome to the show, how are you today?
Kayleigh - I'm good, thank you. Thanks for having me.
Rene – Awesome! So, before start, can you tell us a little bit about yourself and your background as it relates to quantum computing, physics, computer science and so on?
Kayleigh – Sure. So, I got my bachelor's at Indiana university in 2011. I went onto get a PhD in physics at University of Berkeley, University of California, Berkeley and then I did a Post-doc afterwards and joined Adam computing in march of 2020. So, I've been working here for almost two years and I'm a senior quantum engineer.
Rene – Wow! And like this is a great segue into the first question. Because, you know, often times in the show, we talk a lot about the software stack on quantum computers and like what you can already do with quantum inspired computing and these kind of things. But today, we're going to talk about the hardware, right, and so but you know for the whole audience here to be fully inclusive, can you just give us a high-level overview of all the components that are involved in a real physical quantum computer?
Kayleigh - Yeah certainly. I mean, I can try. So I like to think of it in kind of a few different areas. Every quantum computer needs qubits right and gates. So, we have our platform, our qubit register and with a neutral atom quantum computer, we have a lot of back end hardware to create this qubit register. And to reload this qubit register every tens of quantum operations and then we have our, you know, the hardware that provides our gates. So with a neutral atom quantum computer that's lasers and microwaves and lots of fun you know. Optics, electro optical components, acoustic optical components and then we have a control system which really is the conductor- the timekeeper which synchronizes all of these different parts. And with the neutral atom quantum computer, we take these atoms, we reload our register, perform operations, perform circuits. We can keep them around for a while, but eventually we do have to reload them again. So I guess, that's to speak to how important the control system aspect is for our particular platform. But yeah just qubits gates and then a control system, I would say those are the three main components.
Rene – Gotcha. And basically, after you get these neutral items out of the ground stake, right? You need to cool them and cooling is really a critical part to keep the qubits stable right? And so these systems are typically cooled down to almost zero kelvin or absolute zero basically. And this is performed with a special kind of a technique where also a magnetic fields and lasers are involved correct? Can you help us a little bit to put this into perspective, like, what is the time required for this atomic preparation until you actually can do the computation? Also, what is the kind of precision required, you know, how do you actually cool down atoms with a laser?
Kayleigh – Yeah, it's kind of counter-intuitive, isn't it, because we do start with this chunk of strontium, in our case in an oven. We have to heat the atoms to dissociate them in the solid so we sublimate this chunk of strontium and so we have this hot strontium gas. And so from that point forward the name of the game is to cool the gas, because um you know when atoms are like buzzing around, you can just imagine intuitively things can't be done as precisely like, how are you going to, you know, capture this buzzing atom in one of your optical traps? So, we do rely on a lot of these techniques that have been developed over the last 30-40 years of laser cooling. Quantum mechanics says that light, you know, there's a wave particle duality of light. So, you can think in one aspect of what we're doing is just bombarding the atoms with little photons, little balls of light which have momentum. So, you know if an atom is like buzzing or flying this way and we're just kicking it with light counter to its trajectory, we will eventually slow it down and slowing it down it goes hand in hand with cooling. There are more sophisticated tricks we play where like you said, we have magnetic fields to spatially couple a trapping you know a location for trapping the atoms because just shining a laser beam at these hot atoms doesn't really say anything about space, right? I mean there there's no notion of you know coupling this cooling to a spatial location and it does have to be done very precisely, you know, we're talking about hundreds of milliseconds, maybe a little bit longer of this sort of state preparation and you know trapping these atoms to a region at the end within you know millimeters, micrometers because eventually we want to load them into an array of traps where the traps are spaced by micrometers. So, it does have to be done exceptionally well but these tricks and techniques have been pioneered in the field and we're really you know just kind of following along in the footsteps of other academic groups.
Rene – So you just make it sound as it's so easy. But it’s not. It's very much state of the art what you're doing there and well impressive. Thank you for explaining it so nicely. Now I understand it, so like you said like bombarding, well laser is emitting photons right so you're sending these photons into the direction and basically slowing them down which in the end, you know, results in cooling. So nice explain, thank you. And now it makes sense for me as well.
Kayleigh - I should say at the end, we go from 400 degrees celsius to 10 micro kelvin. So we're talking about ten to the seven, what's that, ten to the seven, ten to the seven like reduction in this, like temperature value. So, it's, I mean thinking about it, it's quite impressive, you know that people have figured out how to do this.
Rene - And it's just in a few hundred milliseconds you said, right?
Kayleigh – Yeah.
Rene – Wow. So much so little time to actually fix any arrows and things like that so, yeah precision is really really highly required also in terms of like the whole apparatus. Like you got to make sure if you make little changes you have to probably, try out a few things and a lot of trial and error, I guess.
Kayleigh – Absolutely and that's what I should say i've spent my PhD in postdoc studying is working on these platforms and the quantum engineers here, I mean, that's kind of our specialty is building these sorts of systems in this sort of way and doing it quickly right?
Rene - And well yeah and another really amazing part about the neutral atom approach was when I talked with your colleagues, Denise Ruffner and Robin Coxe in previous QuBites episodes. You can keep those much longer stable, right? She mentioned, like 10 seconds or something like this, where you can keep the qubits in a stable state which is very impressive if you compare it to other approaches like with semi-conductor or trapped ions and so on. It's just what is it nanoseconds milliseconds or something like this right, where they can keep it stable and so well you can do probably much more with 10 seconds of stable computation.
Kayleigh - Yeah our coherence time is tens of seconds long and I mean, really, that's coming about because we're using you know, we're using neutral atoms. We're using the universe's like two level system. We're not using a fabricated two level system that you know someone made. It really is as perfect as it can get. There are subtleties there, but, yeah!
Rene - So, looking at your background, of course, you're not in a lab right now, beautiful pictures there, actually or like decoration. So can you tell us since you're not in the lab but what does a typical work day for you look like and are you mostly actually working in the lab on actual hardware or more on a computer and doing computations and simulations and figuring out certain things? Also, of course, what are the big challenges actually and from scaling to from a more like, say scientific physical system like an apparatus in the end into this commercial quantum computer. I think it's called phoenix, right for Atom computing.
Kayleigh – Yes, our first our first-generation computer is Phoenix which is in Berkeley, what I'm working on. And a typical day can be all of the above you know. I'm asked this question kind of what you do as a quantum engineer and it always kind of stalls me a bit because I work on software. I work on hardware. At the end of the day, I'm one of the people that see the physical system we have and are working to abstract it to the powerful quantum computer. We know it can be so it does require you know work in automation and calibration. Work on this process of cooling to make it better and more robust. Working on our single qubit and two qubit gates to make the fidelities higher, increasing the atoms, we're trapping. So, you kind of do many different things and I think that's why I enjoy this job so much because it's never boring and you get to learn a bunch of different skills. But it's hard to quantify or describe. I do spend a lot of my time in lab wrangling lasers certainly. One of the challenges of making our system commercial right is to remove the increase the uptime and remove kind of the human intervention, that atomic physicists need to do so that's always in mind, how can we make this more hands-off, how can we do this better or add machine learning or you know have some sort of a calibration routine and I think that we really got, you know, when, so I started in march of 2020. I had three days in the office before there was a shelter in place and I was able to watch the other people, the more senior people at the company. We had few remote utilities, we still needed somebody to come into the lab and turn on the laser but say, may of that year you could trap atoms like from your bedroom if you wanted to. So Covid has been challenging but it's been really interesting to see the company's approach to it and you know and ways I think forced us to really get a handle on the things that we need to have to be commercial and to build these machines well. So this was not the case in my post-doc lab. You definitely had to be there for every second of the experiment.
Rene – Gotcha. That's pretty impressive and one way to say is like COVID of course accelerated a lot of remote work but now it also has a related like remote quantum computing labs kind of a thing right. And it's the first and it's the first step to get to the kind of qaas, the quantum as a service, right where you can then bring this phoenix quantum computer into well some place and then you can schedule a job remotely via whatever quantum as a service provider and then it will run on the actual hardware but you don't have to worry about it. You don't have to tune it and this perfectly, like you said, is like the goal is to make this happen, right, to that you don't need a lot of human intervention and things and COVID might have helped well a little bit with that well. We want to see some positive things right? This is awesome. Thank you so much Kayleigh for all this explanation and the insights you shared today. Unfortunately we're already a little bit at the end of the show but we could talk for much longer and I would love to learn much more but again thank you so much for joining us today and explaining it so well. This is very much appreciated.
Kayleigh - Thank you for having me. This was fun.
Rene - Thanks everyone for joining us for yet another episodes of QuBites, your bite-sized pieces of quantum computing. Well, watch our blog and follow our social media channels to hear all about the next episodes and of course visit our website to watch all the previous episodes from season one to five, where we also had other colleagues from Atom computing, where we talked about a couple of other things. So you might want to see those as well. But anyhow, take care and see you soon, bye!
