Sebastian Hassinger (00:00)
Hi, Arneb. Thank you very much for joining me.
Arnab Banerjee (00:04)
Thank you, it's my pleasure to join you here.
Sebastian Hassinger (00:07)
Yeah, it's good to see you again. ⁓ You've got a really interesting story. So before we jump into the current research, the current topic that I want to talk about, can you share a little bit about your path to Purdue and the Quantum Science Center, how you got to where you are today?
Arnab Banerjee (00:21)
Absolutely. So ⁓ I am an experimentalist ⁓ in my DNA. So I do experiments. I grow materials and then I go and measure them ⁓ using cutting edge techniques, which includes neutron scattering, which includes transport, which includes x-ray scattering in many of the national laboratory facilities. So the way I got here is that I did my PhD at University of Chicago.
Sebastian Hassinger (00:28)
Mm-hmm.
Arnab Banerjee (00:52)
with Tom Rosenbaum, who right now, by the way, is the president at Caltech. And he was a really big proponent of quantum computing himself back in the days, back in 2000s. And with him, I did work on classical and quantum phase transitions in magnetic materials, but also got interested in magnetic materials and what kind of magnetic materials could really become.
Sebastian Hassinger (00:52)
Mm-hmm.
Sebastian Hassinger (01:04)
Mm.
Arnab Banerjee (01:20)
⁓ tomorrow's ⁓ quantum materials ⁓ for applications for computing for sensors. that ⁓ research really took me into Oak Ridge National Lab where I was a postdoc. worked on developing and discovering a new kind of a quantum spin liquid ⁓ material called the Kit Hive Quantum Spin Liquid Material.
Sebastian Hassinger (01:28)
Hmm.
Arnab Banerjee (01:47)
And I took a lot of neutron data on that. And what I figured out is that ⁓ the community really lacked a good understanding of how to deal with such complex data sets. ⁓ Things are quantum, but they are using classical tools and classical approximations. ⁓ And because of this dichotomy, ⁓ this problem never gets truly solved. And when the problem never gets truly solved, you can't really ⁓ develop it. ⁓
Arnab Banerjee (02:17)
anymore. So that is what got me into quantum computing because I thought that that would be the way ⁓ in which we can solve these problems and it was that frustration or that pain point which got me into quantum computing in the first place. ⁓ And then I started working in Oak Ridge National Lab with folks like Travis Humble in their quantum computing institute ⁓ and started learning more and more about how can we actually ⁓ use quantum computers to get access to real data sets.
Sebastian Hassinger (02:25)
Hmm.
Arnab Banerjee (02:46)
⁓ And ⁓ then I came to Purdue partly because I think that it has a great quantum computing ⁓ environment. Microsoft Q Station is here. They ⁓ also have great national lab partnerships. And I can actually develop this with students, ⁓ which also ⁓ makes the next generation come up. ⁓ And then ⁓ one thing led to the other. I think that ⁓ we started working with I.
Sebastian Hassinger (02:46)
Hmm.
Sebastian Hassinger (03:04)
Right.
Arnab Banerjee (03:14)
⁓ many companies ⁓ and ⁓ now we are here. So I do experiments ⁓ and try to get data which we could possibly explain with quantum computers. ⁓ And that was my brief history.
Sebastian Hassinger (03:27)
Right, Yeah, yeah, yeah. So you call yourself an experiment through and through. And I think I saw a quote from you I actually think I've heard you say something like this. ⁓ Using words, something like this, like making uncertainty your friend ⁓ is that sort of like how you see ⁓ the challenge of experimentalism, I guess, in physics. ⁓ So you mentioned Kateyev. ⁓ That's the theoretical work.
Sebastian Hassinger (03:56)
that you have been sort of trying to experimentally test and prove out essentially, right? The topological states of matter.
Arnab Banerjee (04:09)
Correct. So, yes, I think that almost the entire electronic industry is built around trying to avoid quantum effects as much as possible. And I think that this is the time when we need to basically make quantum our friend instead of our enemy and fundamentally change the way we can actually miniaturize things.
Sebastian Hassinger (04:10)
Yeah.
Sebastian Hassinger (04:19)
Hmm. ⁓
Sebastian Hassinger (04:29)
Right.
Arnab Banerjee (04:36)
get across that barrier into a new kind of Moore's law, ⁓ which would be over-accomplishing quantum. ⁓ And which means that we need to rethink the way we design electronics. We need to rethink the way we design any circuitry. ⁓ And one of the things that we thought, which is what attracted me to the Kittav magnetism, is because Kittav ⁓ gave this theoretical framework of how to make
Sebastian Hassinger (04:48)
Hmm.
Arnab Banerjee (05:03)
a quantum computer or quantum gates, which are naturally protected from de-coherence. As you know, de-coherence is a big problem. ⁓ And they are naturally protected ⁓ from de-coherence ⁓ by nature. ⁓ So there are materials where nature ⁓ works in a way that ⁓ they are naturally protected. And if you can actually ⁓ take some of these materials or build some of these materials out and
Sebastian Hassinger (05:09)
Right.
Sebastian Hassinger (05:25)
Hmm.
Arnab Banerjee (05:30)
make circuits out of that, then you are already setting yourself up for advantage. Not only that, ⁓ are, ⁓ mean, a material platform is ultimately the most scalable platform as we know from the semiconductor boom. ⁓ So, ⁓ the ⁓ reason I got very interested in that research and I'm still very interested is because ⁓ I think that ⁓ that's one pathway we can get to truly protected qubits which are
Sebastian Hassinger (05:35)
Right.
Sebastian Hassinger (05:44)
Hm
Arnab Banerjee (05:59)
scalable to the billions ⁓ one day. ⁓
Sebastian Hassinger (06:01)
Hmm. Hmm. That's really interesting. And so so that really that pursuit of topological states of exploring topological states led you into ⁓ these, ⁓ this set of or this decade long set of experiments, I would say, right. ⁓ And ⁓ for people who like myself are not condensed matter physicists, you mentioned ⁓ quantum spin fluid, liquid as one of the things that you
Sebastian Hassinger (06:29)
have discovered through your experimentation, what is a quantum spin liquid?
Arnab Banerjee (06:35)
Uh-huh, so ⁓ a quantum spin liquid, no, no, absolutely no problem. ⁓ I've been asked this. ⁓ So ⁓ think of a normal liquid like water. In water, what happens is that all the molecules are randomly moving around because of something called Brownian motion. There seems to be like no rhyme or rhythm. On the other hand,
Sebastian Hassinger (06:36)
⁓ In simple terms, no problem, right, R &F? ⁓
Sebastian Hassinger (06:59)
Mm-hmm.
Arnab Banerjee (07:03)
On a surface of water, you go and throw a stone, ⁓ you form ripples. ⁓ You form this nice vortices, which means that there is an overall orderliness in the way the molecules collectively move, although individual molecules are all moving around like random. That is exactly what you have in a quantum spin liquid, where the spin directions ⁓ basically move collectively.
Sebastian Hassinger (07:12)
Mm.
Arnab Banerjee (07:33)
in ways where they make these really nice dancing patterns which look extremely ordered. But although the spins themselves, ⁓ if you take a snapshot, feel that they are all completely random. ⁓ And from that orderly motions, essentially new kinds of particles can emerge, new kinds of statistics can emerge.
Arnab Banerjee (07:56)
And some of these statistics are topological. Some of these statistics have very interesting patterns, just like in water you see vortices. ⁓ When a ship goes, you see this nice vortices form, and these vortices are really long-lived. That's a really interesting example of a topological state in ⁓ water, in a liquid, water.
Sebastian Hassinger (08:16)
⁓ Right.
Arnab Banerjee (08:18)
⁓ And that's exactly the same kind of a state that you can have in a spin system where these vortices become really long lived in a quantum spin liquid. And if you can encode your information in that, that becomes really long lived and protected. So that's a quantum. ⁓
Sebastian Hassinger (08:31)
Hmm. Right. So intuitively, it sounds like it what you're saying is almost the the analogy here is the topological system ⁓ is like a vortex in the water, which, which has staying powder and power even in the random chaos of the individual molecules of movement ⁓ of water. ⁓ The vortex itself has its own, the spin itself has momentum, it has an actual
Sebastian Hassinger (08:59)
sort of almost lifespan ⁓ while the energy of that vortex runs down. when you say those are observable, do you mean in sort of like the ground state of ⁓ a condensed matter, like a crystalline ⁓ structure or lattice or whatever, that where you're observing?
Arnab Banerjee (09:17)
Yes, so you will need to look at, so there is the electronic ground state ⁓ and beneath the electronic ground state we have the spin ground state. ⁓ So ⁓ once you look at the spin ground state, then you get a very clear signature of these motions. ⁓ And that's what we are.
Sebastian Hassinger (09:27)
Okay.
Sebastian Hassinger (09:34)
⁓ Interesting. ⁓ And so ⁓ you mentioned in ⁓ one of your tools in experimentation was ⁓ neutron diffusion ⁓ and using ⁓ the Oak Ridge ⁓ spallation facility. Is that right? They've got that. So ⁓ how does that work as a tool in your experimentation?
Arnab Banerjee (09:48)
deflection.
Arnab Banerjee (09:57)
That's correct.
Arnab Banerjee (10:03)
⁓ so ⁓ it's extremely interesting. And I can ⁓ speak days about it. ⁓ So ⁓ literally. ⁓ So ⁓ the way a neutron is produced is by having some kind of a material like tungsten or mercury, which has a lot of neutrons. It's a heavy metal. And then bombarding it with very fast protons. And that proton, when it hits this metal, it basically gives out a burst of neutrons.
Sebastian Hassinger (10:32)
Hmm.
Arnab Banerjee (10:32)
What you do is that you slow those neutrons down and then you put them in a tube like a wave guide and then you make them project onto a certain material of your choice which you are trying to study. ⁓ Neutrons have ⁓ three very interesting properties that makes it ⁓ useful. The first property is that the neutrons, ⁓ they have a mass, ⁓ which means that they have something called a de-borigly wavelength. ⁓
Arnab Banerjee (11:01)
in quantum physics terms. In other words, it behaves both like a wave and a particle ⁓ at the same time. ⁓ And the wave nature of that ⁓ neutron means that if it hits a material, it will diffract. ⁓ And on the screen, it will produce a pattern that essentially gives you the information of where your atoms ⁓ are. Not only that.
Sebastian Hassinger (11:06)
Hmm, okay.
Sebastian Hassinger (11:24)
Hmm.
Sebastian Hassinger (11:27)
Hmm. Wow.
Arnab Banerjee (11:31)
If you can figure out how much energy this neutron has lost, ⁓ it will also tell you what are the allowed energy levels in the system, which means that it will tell you how these atoms are actually moving. ⁓ So it gives two information at the same time. Now it gets even better. ⁓ Neutrons not only have a de Borger-Lewe wavelength, they also have a spin.
Sebastian Hassinger (11:41)
Hmm.
Sebastian Hassinger (11:46)
Wow.
Arnab Banerjee (11:58)
So they are spin-half particles, which means that the spins of neutrons will directly interact with the spins of your material. And it produces some of the cleanest, cleanest data of spectroscopy that you can ever get in a magnetic material. And that is what we really ⁓ rely to figure out ⁓ what the material is really doing. All the spectroscopic levels completely separated out in energy and momentum.
Sebastian Hassinger (12:00)
Right.
Arnab Banerjee (12:27)
And so ⁓ it's an extremely powerful technique. And since it works ⁓ as a weak scatterer, ⁓ you can do an exact theory ⁓ on the result which comes out. ⁓ And neutrons have a lot more interesting properties. I mean, it also gets absorbed. It penetrates the material depending on the content of its nucleus. So you can do imaging with neutrons. ⁓
Arnab Banerjee (12:54)
You can do fracture imaging with neutrons very nicely, ⁓ something which is done ⁓ quite a bit in Los Alamos National Laboratory. ⁓ Here we are using it to, ⁓ in our research, are really using it to do a momentum-resolved spectroscopy, and there are very few techniques ⁓ that can do that so cleanly, and neutron scattering is clearly ⁓ one of the winners in that technique space.
Sebastian Hassinger (13:03)
Right.
Sebastian Hassinger (13:20)
Interesting, interesting. ⁓ you know, this ⁓ one of the motivations for this line of experimentation, as you said, was sort of trying to ⁓ improve ⁓ the materials that are used for building qubits, for fabricating qubits, try to make something that will have ⁓ more resistance to errors, ⁓ because of topological states. ⁓
Sebastian Hassinger (13:44)
greater coherence times, those sorts of things. But now you've in the last, I mean, actually since we met, ⁓ you've been exploring, ⁓ this, the use of quantum computers to simulate materials. ⁓ so ⁓ what's the, what's the role, know, when you have that neutron scattering, ⁓ technique, what's the role of quantum computing simulation in the, in the material science experimentation that you're doing?
Arnab Banerjee (14:11)
⁓ Mostly ⁓ there is a twofold requirement. The first thing is that there are ⁓ no ⁓ competent classical techniques as we ⁓ scale up in size and scale up in number of dimensions, which means that you only have approximate classical techniques or no classical techniques. ⁓ And then if you have really nice data and nothing to analyze that data with, then that's ⁓
Sebastian Hassinger (14:40)
Hmm.
Arnab Banerjee (14:41)
fail signs ⁓ in some way. mean, and I think that that's what quantum computing can really solve ⁓ is make that connection right away. And then it can predict. ⁓ And exactly nature isn't classical. And since nature isn't classical, we need a quantum tool to solve a quantum problem. And as we see ⁓ that
Sebastian Hassinger (14:44)
Mm-hmm.
Sebastian Hassinger (14:49)
Right.
Sebastian Hassinger (14:53)
Nature isn't classical, as Feynman said. ⁓
Arnab Banerjee (15:11)
⁓ Then we can go and ask these questions, ⁓ which we cannot ask right now very well, by the way, because we cannot solve things. We can start asking these questions that, okay, now we know what the system is really doing. Can we see what is happening microscopically? Can we look at the, can we perturb the Hamiltonian? Can we actually see if this could be a good sensor? Can it be a good qubit? Can it be a good ⁓ information ⁓ carrying bus from
Arnab Banerjee (15:39)
one place to the other. And all those questions we can only ask when we understand the material better. ⁓ And that's what I think it really is. ⁓
Sebastian Hassinger (15:47)
That's great. That's great. And so this recent work, I think was just published to the archive about 10 days ago at the time of this recording. It's a collaboration with Oak Ridge, with Los Alamos, with IBM. And essentially, if I'm correct, you're using the Heron chip or about 50 qubits on the Heron chip. think the Heron is 150 something to essentially simulate or neutron scattering data.
Sebastian Hassinger (16:17)
and see if you can simulate ⁓ accurately enough to match your actual experimental results from real neutron scattering. that right? Yeah. ⁓
Arnab Banerjee (16:25)
Exactly, and this is one step we absolutely need to do before we tackle more complex, ⁓ more complex neutron scattering data.
Sebastian Hassinger (16:31)
Of course. Yeah.
Sebastian Hassinger (16:35)
You have to have a reference point to prove that your simulation actually is valid before you take off into the realm where you can't actually ⁓ simulate it classically or validate it. Yeah. Yeah. Interesting.
Arnab Banerjee (16:42)
⁓ Exactly. Exactly. ⁓ We need to know whether what we are doing really makes sense. And ⁓ that's what this experiment is about.
Sebastian Hassinger (16:50)
Yeah.
Sebastian Hassinger (16:53)
Yeah, yeah. So walk me through the details of the like, what was the experimental data you had? And how did you simulate that on the on the IBM system?
Arnab Banerjee (17:03)
So the experimental data was partly obtained ⁓ at Rutherford Appleton Lab, ⁓ fondly called ISIS by the way, ⁓ that's in the United Kingdom. ⁓ And then partly it was obtained at Oak Ridge National Lab and then it was stitched together. And many of these works were happening over ⁓ quite some time. In fact, this is a very well studied material and
Sebastian Hassinger (17:18)
Mm-hmm.
Arnab Banerjee (17:33)
there is this groups of ⁓ Alan Tennant, Steve Nagler, who had been ⁓ working on this material for a long time. And turns out that this is a really nice Heisenberg spin chain, which complies with, ⁓ which almost complies nicely with Bethe ansatz solution named after Hans Bethe, who first proposed it solution exactly. ⁓ And ⁓ it does produce a type of a quantum spin liquid ⁓ in one dimension.
Arnab Banerjee (18:01)
which is ⁓ very akin to something called a Luttinger liquid, which ⁓ now that's ⁓ a lot of jargons, but ⁓ overall, ⁓ it's a really nice problem and it's the problem which gives us the confidence that, we are getting interesting quantum states ⁓ out ⁓ that we've seen a real material, but now we also see it in ⁓ a quantum computer so that we can make a one-to-one benchmark.
Sebastian Hassinger (18:01)
Hmm.
Sebastian Hassinger (18:20)
Mm-hmm.
Arnab Banerjee (18:29)
And so ⁓ this material basically shows these really nice ⁓ quasi particles called spinons, ⁓ which basically are fractional. What that means is that it is as if the electron has broken apart into fractions of itself, which is not possible normally. But ⁓ in a quasi particle sense, in a many-body system, that is effectively true. ⁓
Sebastian Hassinger (18:29)
right.
Sebastian Hassinger (18:44)
Hmm.
Arnab Banerjee (18:56)
So you see these very interesting properties, and we thought that this would be the material that would convince us that we are ready to tackle complex problems. At least let's show that in one dimension. In one dimension, of course, we are not talking about quantum supremacy per se, but it is a great benchmark problem. ⁓
Sebastian Hassinger (19:12)
Right, Benchmark, yeah, yeah. And ⁓ I'm curious that the reason for selecting the particular ⁓ experiment to... ⁓
Sebastian Hassinger (19:24)
to test your simulation on the IBM system, ⁓ the ⁓ reason for choosing something that had the spin ⁓ systems, spin liquid systems, is that because ⁓ it's a good workout for your simulation, for your circuit, it's a good challenge to sort of stretch, or is it because of your ⁓ pursuit of spin liquids as a potential ⁓ key for unlocking better?
Sebastian Hassinger (19:54)
materials for quantum computing or both. ⁓
Arnab Banerjee (19:57)
⁓ I think that it is ⁓ the first one in this ⁓ case ⁓ because ⁓ you want to see that ⁓ your quantum computer is giving you emergent states. ⁓ Emergent states are states where you have collective excitations. These are states which do not natively exist in the two-body interaction in the qubits, but can emerge
Sebastian Hassinger (20:01)
Okay.
Arnab Banerjee (20:27)
effectively ⁓ as you ⁓ evolve the system. That's when you know that, ⁓ yes, mean, all the spins are doing or all the qubits are doing all its random things, but at the end, it's giving you something which is more than the sum of its parts. ⁓ And this is one problem which helps us benchmark that. And that is what guided us to this problem. And it turns out that many of these type of problems are
Sebastian Hassinger (20:44)
Right.
Arnab Banerjee (20:56)
spin liquid problems, we like it or not, of different kinds. ⁓ And there are non-spin liquid varieties of these problems as well. But ⁓ this is ⁓ a problem space where we have really nice data ⁓ and ⁓ we trust that data. We can benchmark it not only against theory, but also against experiments. ⁓ And so that's what really got us
Sebastian Hassinger (20:58)
Hmm. Hmm.
Sebastian Hassinger (21:16)
Hmm.
Arnab Banerjee (21:25)
into it. But yeah, I mean, ⁓ we have still not gotten to the types of materials or Hamiltonians that would give us a true quantum advantage or could help us make qubits or sensors or anything. But that's the yeah.
Sebastian Hassinger (21:34)
Right.
Sebastian Hassinger (21:37)
Right, right. And that when you refer to a quantum advantage, in your mind is that by definition a topological qubit of Majorana, a fermion sort of based qubit, or do you think spin liquids may afford other advantages in the characteristics of the materials, longer coherence times or other means of providing a better foundation for fabricating qubits?
Arnab Banerjee (22:03)
It may, so ⁓ this is basically research that will, some of it will still need to be done. There are actually frameworks for quantum spin liquid based qubits. ⁓ One of them was just published in physical review X by Jason Alicia from Caltech. ⁓ So ⁓ there are frameworks for that. Now ⁓ the engineering towards that will now need to catch up of course.
Sebastian Hassinger (22:15)
Hmm.
Sebastian Hassinger (22:32)
⁓ As is always the case. ⁓
Arnab Banerjee (22:32)
⁓ And ⁓ as is always the case. ⁓ And what I am hoping is that we also need to understand not only the engineering, we also need to have enough confidence on the materials to pour in, let's say, ⁓ a couple of hundred of million dollars on that research so that it develops. And I think that what ⁓ the quantum computing research can help is to provide that level of confidence.
Sebastian Hassinger (22:51)
Right, Yeah. ⁓
Arnab Banerjee (23:00)
that modicum of confidence that actually brings in ⁓ that ⁓ funding into the space to develop ⁓ these new systems. having seen that, Marana qubits, as you know, are being pursued by ⁓ Microsoft ⁓ with varying degree of success ⁓ and ⁓ protected qubits ⁓ are being pursued by everybody, by Continuum, by Google.
Sebastian Hassinger (23:00)
Hmm.
Sebastian Hassinger (23:05)
Right.
Sebastian Hassinger (23:10)
Right.
Sebastian Hassinger (23:16)
Mm-hmm.
Sebastian Hassinger (23:29)
Mm-hmm. Yeah.
Arnab Banerjee (23:31)
All of these are actually some type of a surface code, ⁓ some kind of a topological code which they are employed.
Sebastian Hassinger (23:35)
Hmm. Right. That is fascinating to me that, that, ⁓ surface code is, is just a logical set of operations that are in a sense, simulating a topological state of matter. And, potentially you could achieve a topological state of matter to instantiate a qubit physically and have that protected logic sort of embedded in the, in the design essentially of the, of the material. And then the qubit that's.
Arnab Banerjee (23:51)
Exactly.
Sebastian Hassinger (24:05)
realized in the material. It's so fascinating. I love that feedback loop of ⁓ quantum computers helping with quantum physics experimentation and research, which then helps making better quantum computers. ⁓ When we first met, you were doing a simulation in collaboration with Cuera on the Aquila device. was a Shastri Sutherland material simulation. You've done work with the
Arnab Banerjee (24:06)
Exactly.
Sebastian Hassinger (24:33)
⁓ D wave quantum annealing system as well. Now you're working with IBM system Do you see sort of the the modalities as all having different roles to play at this stage and in the in the experimentation?
Arnab Banerjee (24:47)
This is actually a very important question and ⁓ I have been asked by many people ⁓ and actually I am ⁓ one of the groups who has worked with ⁓ many of these modalities ⁓ and ⁓ each back-end has its ⁓ each modality has its advantages and disadvantages. I'll give you an example ⁓ IBM has qubits which are extremely fast they are transplant based programmable
Sebastian Hassinger (24:58)
Mm-hmm.
Arnab Banerjee (25:16)
and ⁓ but they have a 2D lattice. ⁓ For the research that we are doing, it actually requires many, many quick, cheap qubits. So that's a bit fantastic. That's ⁓ the operation. So that's fantastic for the research. So that's a great fit. ⁓ Something like Continuum, for example, has a much more all-to-all connectivity, but they are slower. ⁓ So if somebody really needs that all-to-all connection for some...
Sebastian Hassinger (25:27)
Mm. Mm.
Arnab Banerjee (25:44)
kind of Hamiltonian, then they will better be using that. On the other hand, if somebody says that, I don't need so much of programmability of ⁓ XXY, I don't need to measure in XYZ, all kinds of bases, ⁓ Heisenberg is not required. I only need an effective model, which is Ising-like and a transverse field Ising is gonna be, make good for me. Then I think that,
Arnab Banerjee (26:13)
D-Wave and Q-ERA will, ⁓ depending on how quantum you need to be ⁓ in the overall scale and how quick you need to be, I think that they offer ⁓ better products. So it really depends on the problem that we are trying to solve ⁓ and ⁓ we really need to fit the ⁓ backend modality to the problem. And this is where I think the success can really happen.
Sebastian Hassinger (26:30)
Hmm.
Sebastian Hassinger (26:36)
Mm-hmm.
Sebastian Hassinger (26:40)
That's really interesting. You've written a roadmap paper on materials for quantum technologies. What do you think the pipeline looks like from your perspective, having done these experiments on a variety of different modalities and being very close to obviously coming from the physical experimentation and now doing simulation? How far are we from
Sebastian Hassinger (27:04)
discovering or improving materials ⁓ for direct use in the design and fabrication of qubits do you think?
Arnab Banerjee (27:12)
⁓ so I think that as far as quantum ⁓ helping is concerned, we are actually quite close. will, I'm actually ⁓ monitoring this space of going into error corrected qubits or error corrected logical qubits. I mean, I am both optimistic and ⁓ cautious at the same time because ⁓ optimistic because it will probably be better qubits, but ⁓ cautious because the effective number of qubits will shrink.
Sebastian Hassinger (27:34)
Of course. ⁓
Arnab Banerjee (27:43)
and the effective ⁓ kind of operations that you can do with those qubits will also probably change. So ⁓ I think that ⁓ the trajectory of this research is going to change over the next two, three years as we go into that space. ⁓ And ⁓ it will be interesting to see where we land at the other side of it. Having said that, ⁓ I think that ⁓
Sebastian Hassinger (27:43)
Right.
Sebastian Hassinger (27:59)
Hmm.
Arnab Banerjee (28:09)
⁓ The goal is to basically make sure that maybe 70 or 80 or 90 qubits ⁓ in some kind of a 2D geometry ⁓ is ⁓ coherently coupled for ⁓ an amount of time, ⁓ which is probably like J by 20, where J ⁓ is the interaction scale. And then we are already getting into a point where
Sebastian Hassinger (28:34)
Hmm.
Arnab Banerjee (28:38)
there are classical techniques that will not be able to do it. So from that standpoint, I think that we are just a factor of do-off, if we need to.
Sebastian Hassinger (28:42)
Mm-hmm.
Sebastian Hassinger (28:48)
Right. And sorry, do you mean 70 to 90 cubits, ⁓ logical cubits that are fully protected with surface code? Or do you mean sort of good enough cubits ⁓ with enough connectivity? Yeah. Yeah. Yeah. So maybe, maybe error mitigation or high enough fidelity for the two cubic gates ⁓ and enough connectivity. Yeah. Yeah. That's, that's not that far off. ⁓ That's interesting. That's really interesting.
Arnab Banerjee (28:58)
good enough cubits good enough cubits
Arnab Banerjee (29:05)
⁓ Correct, correct, ⁓ And from that standpoint, yeah. Exactly. Yeah. ⁓ The choice of problems would be ⁓ important though ⁓ in that space.
Sebastian Hassinger (29:21)
Of course.
Sebastian Hassinger (29:23)
Yeah, I mean, I think that's where your set of experiences becomes really, really crucial because you can draw on your experimental experience and tailor the experiment to the characteristics of that system to choose a problem that is not classically accessible, but is accessible to that particular systems architecture and characteristics. So that's really exciting. ⁓ so, okay, last question you've, you know, with, with the
Sebastian Hassinger (29:51)
years of physical experimentation with neutron scattering. we, you know, forget about 70 to 90 good enough qubits. If you say, okay, you've got a quantum computer of arbitrary quality and scale. What's the material science of the neutron scattering problem you would throw that system at first? What's the answer you want to get overall?
Arnab Banerjee (30:12)
⁓ Well, there are quite so many, ⁓ I think that ⁓ I would ⁓ like to simulate the entire standard model using a ⁓ quantum computer.
Sebastian Hassinger (30:17)
Yeah. ⁓
Arnab Banerjee (30:35)
Get everything. ⁓
Sebastian Hassinger (30:35)
my that's a modest goal. Yeah. ⁓ Yeah, theory of everything. Okay. Yeah, I think you're you're not alone in that. I think that's what everybody but so that would be but that would be from the perspective of starting with material science and building up essentially, right? You would be
Arnab Banerjee (30:54)
⁓ Right, I mean more realistically, I think that there are problems which are completely untractable using classical computers that we can tackle. I mean the frustrated triangular lattice, it's so simple to think about it, but it's so inaccessible. ⁓ And ⁓ any model where you have ⁓ more connectivity, the more the connectivity you have the the quantum. ⁓
Sebastian Hassinger (31:10)
Hmm.
Sebastian Hassinger (31:16)
Right.
Arnab Banerjee (31:20)
entanglement kind of spread so fast that you are completely clueless about what is happening after ⁓ a certain amount of time. ⁓ And any of these problems become quite accessible. Understanding what's the interaction between spins and electrons, ⁓ it's ⁓ mostly ⁓ a problem ⁓ which ⁓ many theorists don't want to touch.
Sebastian Hassinger (31:37)
Hmm.
Arnab Banerjee (31:45)
And I think that ⁓ those will start getting accessible and we will get a much better understanding of what's going on around us.
Sebastian Hassinger (31:45)
Yeah.
Sebastian Hassinger (31:54)
Mm That's great. I mean, I think, like I said, I think your set of experiences makes you the type of person who's uniquely able to navigate this, this, you know, emerging space, you know, as we get closer and closer to being able to tackle classic, you know, in track, classically intractable problems, I think.
Sebastian Hassinger (32:16)
I'll be watching you because I think you're going to be among the first to get to those really interesting results. ⁓ So yet another really nice piece of work. really, I'm always quite impressed with the work you're doing. So thank you so much for joining me today.
Arnab Banerjee (32:30)
Thank you, it was my pleasure.
Hi, Arneb. Thank you very much for joining me.
Arnab Banerjee (00:04)
Thank you, it's my pleasure to join you here.
Sebastian Hassinger (00:07)
Yeah, it's good to see you again. ⁓ You've got a really interesting story. So before we jump into the current research, the current topic that I want to talk about, can you share a little bit about your path to Purdue and the Quantum Science Center, how you got to where you are today?
Arnab Banerjee (00:21)
Absolutely. So ⁓ I am an experimentalist ⁓ in my DNA. So I do experiments. I grow materials and then I go and measure them ⁓ using cutting edge techniques, which includes neutron scattering, which includes transport, which includes x-ray scattering in many of the national laboratory facilities. So the way I got here is that I did my PhD at University of Chicago.
Sebastian Hassinger (00:28)
Mm-hmm.
Arnab Banerjee (00:52)
with Tom Rosenbaum, who right now, by the way, is the president at Caltech. And he was a really big proponent of quantum computing himself back in the days, back in 2000s. And with him, I did work on classical and quantum phase transitions in magnetic materials, but also got interested in magnetic materials and what kind of magnetic materials could really become.
Sebastian Hassinger (00:52)
Mm-hmm.
Sebastian Hassinger (01:04)
Mm.
Arnab Banerjee (01:20)
⁓ tomorrow's ⁓ quantum materials ⁓ for applications for computing for sensors. that ⁓ research really took me into Oak Ridge National Lab where I was a postdoc. worked on developing and discovering a new kind of a quantum spin liquid ⁓ material called the Kit Hive Quantum Spin Liquid Material.
Sebastian Hassinger (01:28)
Hmm.
Arnab Banerjee (01:47)
And I took a lot of neutron data on that. And what I figured out is that ⁓ the community really lacked a good understanding of how to deal with such complex data sets. ⁓ Things are quantum, but they are using classical tools and classical approximations. ⁓ And because of this dichotomy, ⁓ this problem never gets truly solved. And when the problem never gets truly solved, you can't really ⁓ develop it. ⁓
Arnab Banerjee (02:17)
anymore. So that is what got me into quantum computing because I thought that that would be the way ⁓ in which we can solve these problems and it was that frustration or that pain point which got me into quantum computing in the first place. ⁓ And then I started working in Oak Ridge National Lab with folks like Travis Humble in their quantum computing institute ⁓ and started learning more and more about how can we actually ⁓ use quantum computers to get access to real data sets.
Sebastian Hassinger (02:25)
Hmm.
Arnab Banerjee (02:46)
⁓ And ⁓ then I came to Purdue partly because I think that it has a great quantum computing ⁓ environment. Microsoft Q Station is here. They ⁓ also have great national lab partnerships. And I can actually develop this with students, ⁓ which also ⁓ makes the next generation come up. ⁓ And then ⁓ one thing led to the other. I think that ⁓ we started working with I.
Sebastian Hassinger (02:46)
Hmm.
Sebastian Hassinger (03:04)
Right.
Arnab Banerjee (03:14)
⁓ many companies ⁓ and ⁓ now we are here. So I do experiments ⁓ and try to get data which we could possibly explain with quantum computers. ⁓ And that was my brief history.
Sebastian Hassinger (03:27)
Right, Yeah, yeah, yeah. So you call yourself an experiment through and through. And I think I saw a quote from you I actually think I've heard you say something like this. ⁓ Using words, something like this, like making uncertainty your friend ⁓ is that sort of like how you see ⁓ the challenge of experimentalism, I guess, in physics. ⁓ So you mentioned Kateyev. ⁓ That's the theoretical work.
Sebastian Hassinger (03:56)
that you have been sort of trying to experimentally test and prove out essentially, right? The topological states of matter.
Arnab Banerjee (04:09)
Correct. So, yes, I think that almost the entire electronic industry is built around trying to avoid quantum effects as much as possible. And I think that this is the time when we need to basically make quantum our friend instead of our enemy and fundamentally change the way we can actually miniaturize things.
Sebastian Hassinger (04:10)
Yeah.
Sebastian Hassinger (04:19)
Hmm. ⁓
Sebastian Hassinger (04:29)
Right.
Arnab Banerjee (04:36)
get across that barrier into a new kind of Moore's law, ⁓ which would be over-accomplishing quantum. ⁓ And which means that we need to rethink the way we design electronics. We need to rethink the way we design any circuitry. ⁓ And one of the things that we thought, which is what attracted me to the Kittav magnetism, is because Kittav ⁓ gave this theoretical framework of how to make
Sebastian Hassinger (04:48)
Hmm.
Arnab Banerjee (05:03)
a quantum computer or quantum gates, which are naturally protected from de-coherence. As you know, de-coherence is a big problem. ⁓ And they are naturally protected ⁓ from de-coherence ⁓ by nature. ⁓ So there are materials where nature ⁓ works in a way that ⁓ they are naturally protected. And if you can actually ⁓ take some of these materials or build some of these materials out and
Sebastian Hassinger (05:09)
Right.
Sebastian Hassinger (05:25)
Hmm.
Arnab Banerjee (05:30)
make circuits out of that, then you are already setting yourself up for advantage. Not only that, ⁓ are, ⁓ mean, a material platform is ultimately the most scalable platform as we know from the semiconductor boom. ⁓ So, ⁓ the ⁓ reason I got very interested in that research and I'm still very interested is because ⁓ I think that ⁓ that's one pathway we can get to truly protected qubits which are
Sebastian Hassinger (05:35)
Right.
Sebastian Hassinger (05:44)
Hm
Arnab Banerjee (05:59)
scalable to the billions ⁓ one day. ⁓
Sebastian Hassinger (06:01)
Hmm. Hmm. That's really interesting. And so so that really that pursuit of topological states of exploring topological states led you into ⁓ these, ⁓ this set of or this decade long set of experiments, I would say, right. ⁓ And ⁓ for people who like myself are not condensed matter physicists, you mentioned ⁓ quantum spin fluid, liquid as one of the things that you
Sebastian Hassinger (06:29)
have discovered through your experimentation, what is a quantum spin liquid?
Arnab Banerjee (06:35)
Uh-huh, so ⁓ a quantum spin liquid, no, no, absolutely no problem. ⁓ I've been asked this. ⁓ So ⁓ think of a normal liquid like water. In water, what happens is that all the molecules are randomly moving around because of something called Brownian motion. There seems to be like no rhyme or rhythm. On the other hand,
Sebastian Hassinger (06:36)
⁓ In simple terms, no problem, right, R &F? ⁓
Sebastian Hassinger (06:59)
Mm-hmm.
Arnab Banerjee (07:03)
On a surface of water, you go and throw a stone, ⁓ you form ripples. ⁓ You form this nice vortices, which means that there is an overall orderliness in the way the molecules collectively move, although individual molecules are all moving around like random. That is exactly what you have in a quantum spin liquid, where the spin directions ⁓ basically move collectively.
Sebastian Hassinger (07:12)
Mm.
Arnab Banerjee (07:33)
in ways where they make these really nice dancing patterns which look extremely ordered. But although the spins themselves, ⁓ if you take a snapshot, feel that they are all completely random. ⁓ And from that orderly motions, essentially new kinds of particles can emerge, new kinds of statistics can emerge.
Arnab Banerjee (07:56)
And some of these statistics are topological. Some of these statistics have very interesting patterns, just like in water you see vortices. ⁓ When a ship goes, you see this nice vortices form, and these vortices are really long-lived. That's a really interesting example of a topological state in ⁓ water, in a liquid, water.
Sebastian Hassinger (08:16)
⁓ Right.
Arnab Banerjee (08:18)
⁓ And that's exactly the same kind of a state that you can have in a spin system where these vortices become really long lived in a quantum spin liquid. And if you can encode your information in that, that becomes really long lived and protected. So that's a quantum. ⁓
Sebastian Hassinger (08:31)
Hmm. Right. So intuitively, it sounds like it what you're saying is almost the the analogy here is the topological system ⁓ is like a vortex in the water, which, which has staying powder and power even in the random chaos of the individual molecules of movement ⁓ of water. ⁓ The vortex itself has its own, the spin itself has momentum, it has an actual
Sebastian Hassinger (08:59)
sort of almost lifespan ⁓ while the energy of that vortex runs down. when you say those are observable, do you mean in sort of like the ground state of ⁓ a condensed matter, like a crystalline ⁓ structure or lattice or whatever, that where you're observing?
Arnab Banerjee (09:17)
Yes, so you will need to look at, so there is the electronic ground state ⁓ and beneath the electronic ground state we have the spin ground state. ⁓ So ⁓ once you look at the spin ground state, then you get a very clear signature of these motions. ⁓ And that's what we are.
Sebastian Hassinger (09:27)
Okay.
Sebastian Hassinger (09:34)
⁓ Interesting. ⁓ And so ⁓ you mentioned in ⁓ one of your tools in experimentation was ⁓ neutron diffusion ⁓ and using ⁓ the Oak Ridge ⁓ spallation facility. Is that right? They've got that. So ⁓ how does that work as a tool in your experimentation?
Arnab Banerjee (09:48)
deflection.
Arnab Banerjee (09:57)
That's correct.
Arnab Banerjee (10:03)
⁓ so ⁓ it's extremely interesting. And I can ⁓ speak days about it. ⁓ So ⁓ literally. ⁓ So ⁓ the way a neutron is produced is by having some kind of a material like tungsten or mercury, which has a lot of neutrons. It's a heavy metal. And then bombarding it with very fast protons. And that proton, when it hits this metal, it basically gives out a burst of neutrons.
Sebastian Hassinger (10:32)
Hmm.
Arnab Banerjee (10:32)
What you do is that you slow those neutrons down and then you put them in a tube like a wave guide and then you make them project onto a certain material of your choice which you are trying to study. ⁓ Neutrons have ⁓ three very interesting properties that makes it ⁓ useful. The first property is that the neutrons, ⁓ they have a mass, ⁓ which means that they have something called a de-borigly wavelength. ⁓
Arnab Banerjee (11:01)
in quantum physics terms. In other words, it behaves both like a wave and a particle ⁓ at the same time. ⁓ And the wave nature of that ⁓ neutron means that if it hits a material, it will diffract. ⁓ And on the screen, it will produce a pattern that essentially gives you the information of where your atoms ⁓ are. Not only that.
Sebastian Hassinger (11:06)
Hmm, okay.
Sebastian Hassinger (11:24)
Hmm.
Sebastian Hassinger (11:27)
Hmm. Wow.
Arnab Banerjee (11:31)
If you can figure out how much energy this neutron has lost, ⁓ it will also tell you what are the allowed energy levels in the system, which means that it will tell you how these atoms are actually moving. ⁓ So it gives two information at the same time. Now it gets even better. ⁓ Neutrons not only have a de Borger-Lewe wavelength, they also have a spin.
Sebastian Hassinger (11:41)
Hmm.
Sebastian Hassinger (11:46)
Wow.
Arnab Banerjee (11:58)
So they are spin-half particles, which means that the spins of neutrons will directly interact with the spins of your material. And it produces some of the cleanest, cleanest data of spectroscopy that you can ever get in a magnetic material. And that is what we really ⁓ rely to figure out ⁓ what the material is really doing. All the spectroscopic levels completely separated out in energy and momentum.
Sebastian Hassinger (12:00)
Right.
Arnab Banerjee (12:27)
And so ⁓ it's an extremely powerful technique. And since it works ⁓ as a weak scatterer, ⁓ you can do an exact theory ⁓ on the result which comes out. ⁓ And neutrons have a lot more interesting properties. I mean, it also gets absorbed. It penetrates the material depending on the content of its nucleus. So you can do imaging with neutrons. ⁓
Arnab Banerjee (12:54)
You can do fracture imaging with neutrons very nicely, ⁓ something which is done ⁓ quite a bit in Los Alamos National Laboratory. ⁓ Here we are using it to, ⁓ in our research, are really using it to do a momentum-resolved spectroscopy, and there are very few techniques ⁓ that can do that so cleanly, and neutron scattering is clearly ⁓ one of the winners in that technique space.
Sebastian Hassinger (13:03)
Right.
Sebastian Hassinger (13:20)
Interesting, interesting. ⁓ you know, this ⁓ one of the motivations for this line of experimentation, as you said, was sort of trying to ⁓ improve ⁓ the materials that are used for building qubits, for fabricating qubits, try to make something that will have ⁓ more resistance to errors, ⁓ because of topological states. ⁓
Sebastian Hassinger (13:44)
greater coherence times, those sorts of things. But now you've in the last, I mean, actually since we met, ⁓ you've been exploring, ⁓ this, the use of quantum computers to simulate materials. ⁓ so ⁓ what's the, what's the role, know, when you have that neutron scattering, ⁓ technique, what's the role of quantum computing simulation in the, in the material science experimentation that you're doing?
Arnab Banerjee (14:11)
⁓ Mostly ⁓ there is a twofold requirement. The first thing is that there are ⁓ no ⁓ competent classical techniques as we ⁓ scale up in size and scale up in number of dimensions, which means that you only have approximate classical techniques or no classical techniques. ⁓ And then if you have really nice data and nothing to analyze that data with, then that's ⁓
Sebastian Hassinger (14:40)
Hmm.
Arnab Banerjee (14:41)
fail signs ⁓ in some way. mean, and I think that that's what quantum computing can really solve ⁓ is make that connection right away. And then it can predict. ⁓ And exactly nature isn't classical. And since nature isn't classical, we need a quantum tool to solve a quantum problem. And as we see ⁓ that
Sebastian Hassinger (14:44)
Mm-hmm.
Sebastian Hassinger (14:49)
Right.
Sebastian Hassinger (14:53)
Nature isn't classical, as Feynman said. ⁓
Arnab Banerjee (15:11)
⁓ Then we can go and ask these questions, ⁓ which we cannot ask right now very well, by the way, because we cannot solve things. We can start asking these questions that, okay, now we know what the system is really doing. Can we see what is happening microscopically? Can we look at the, can we perturb the Hamiltonian? Can we actually see if this could be a good sensor? Can it be a good qubit? Can it be a good ⁓ information ⁓ carrying bus from
Arnab Banerjee (15:39)
one place to the other. And all those questions we can only ask when we understand the material better. ⁓ And that's what I think it really is. ⁓
Sebastian Hassinger (15:47)
That's great. That's great. And so this recent work, I think was just published to the archive about 10 days ago at the time of this recording. It's a collaboration with Oak Ridge, with Los Alamos, with IBM. And essentially, if I'm correct, you're using the Heron chip or about 50 qubits on the Heron chip. think the Heron is 150 something to essentially simulate or neutron scattering data.
Sebastian Hassinger (16:17)
and see if you can simulate ⁓ accurately enough to match your actual experimental results from real neutron scattering. that right? Yeah. ⁓
Arnab Banerjee (16:25)
Exactly, and this is one step we absolutely need to do before we tackle more complex, ⁓ more complex neutron scattering data.
Sebastian Hassinger (16:31)
Of course. Yeah.
Sebastian Hassinger (16:35)
You have to have a reference point to prove that your simulation actually is valid before you take off into the realm where you can't actually ⁓ simulate it classically or validate it. Yeah. Yeah. Interesting.
Arnab Banerjee (16:42)
⁓ Exactly. Exactly. ⁓ We need to know whether what we are doing really makes sense. And ⁓ that's what this experiment is about.
Sebastian Hassinger (16:50)
Yeah.
Sebastian Hassinger (16:53)
Yeah, yeah. So walk me through the details of the like, what was the experimental data you had? And how did you simulate that on the on the IBM system?
Arnab Banerjee (17:03)
So the experimental data was partly obtained ⁓ at Rutherford Appleton Lab, ⁓ fondly called ISIS by the way, ⁓ that's in the United Kingdom. ⁓ And then partly it was obtained at Oak Ridge National Lab and then it was stitched together. And many of these works were happening over ⁓ quite some time. In fact, this is a very well studied material and
Sebastian Hassinger (17:18)
Mm-hmm.
Arnab Banerjee (17:33)
there is this groups of ⁓ Alan Tennant, Steve Nagler, who had been ⁓ working on this material for a long time. And turns out that this is a really nice Heisenberg spin chain, which complies with, ⁓ which almost complies nicely with Bethe ansatz solution named after Hans Bethe, who first proposed it solution exactly. ⁓ And ⁓ it does produce a type of a quantum spin liquid ⁓ in one dimension.
Arnab Banerjee (18:01)
which is ⁓ very akin to something called a Luttinger liquid, which ⁓ now that's ⁓ a lot of jargons, but ⁓ overall, ⁓ it's a really nice problem and it's the problem which gives us the confidence that, we are getting interesting quantum states ⁓ out ⁓ that we've seen a real material, but now we also see it in ⁓ a quantum computer so that we can make a one-to-one benchmark.
Sebastian Hassinger (18:01)
Hmm.
Sebastian Hassinger (18:20)
Mm-hmm.
Arnab Banerjee (18:29)
And so ⁓ this material basically shows these really nice ⁓ quasi particles called spinons, ⁓ which basically are fractional. What that means is that it is as if the electron has broken apart into fractions of itself, which is not possible normally. But ⁓ in a quasi particle sense, in a many-body system, that is effectively true. ⁓
Sebastian Hassinger (18:29)
right.
Sebastian Hassinger (18:44)
Hmm.
Arnab Banerjee (18:56)
So you see these very interesting properties, and we thought that this would be the material that would convince us that we are ready to tackle complex problems. At least let's show that in one dimension. In one dimension, of course, we are not talking about quantum supremacy per se, but it is a great benchmark problem. ⁓
Sebastian Hassinger (19:12)
Right, Benchmark, yeah, yeah. And ⁓ I'm curious that the reason for selecting the particular ⁓ experiment to... ⁓
Sebastian Hassinger (19:24)
to test your simulation on the IBM system, ⁓ the ⁓ reason for choosing something that had the spin ⁓ systems, spin liquid systems, is that because ⁓ it's a good workout for your simulation, for your circuit, it's a good challenge to sort of stretch, or is it because of your ⁓ pursuit of spin liquids as a potential ⁓ key for unlocking better?
Sebastian Hassinger (19:54)
materials for quantum computing or both. ⁓
Arnab Banerjee (19:57)
⁓ I think that it is ⁓ the first one in this ⁓ case ⁓ because ⁓ you want to see that ⁓ your quantum computer is giving you emergent states. ⁓ Emergent states are states where you have collective excitations. These are states which do not natively exist in the two-body interaction in the qubits, but can emerge
Sebastian Hassinger (20:01)
Okay.
Arnab Banerjee (20:27)
effectively ⁓ as you ⁓ evolve the system. That's when you know that, ⁓ yes, mean, all the spins are doing or all the qubits are doing all its random things, but at the end, it's giving you something which is more than the sum of its parts. ⁓ And this is one problem which helps us benchmark that. And that is what guided us to this problem. And it turns out that many of these type of problems are
Sebastian Hassinger (20:44)
Right.
Arnab Banerjee (20:56)
spin liquid problems, we like it or not, of different kinds. ⁓ And there are non-spin liquid varieties of these problems as well. But ⁓ this is ⁓ a problem space where we have really nice data ⁓ and ⁓ we trust that data. We can benchmark it not only against theory, but also against experiments. ⁓ And so that's what really got us
Sebastian Hassinger (20:58)
Hmm. Hmm.
Sebastian Hassinger (21:16)
Hmm.
Arnab Banerjee (21:25)
into it. But yeah, I mean, ⁓ we have still not gotten to the types of materials or Hamiltonians that would give us a true quantum advantage or could help us make qubits or sensors or anything. But that's the yeah.
Sebastian Hassinger (21:34)
Right.
Sebastian Hassinger (21:37)
Right, right. And that when you refer to a quantum advantage, in your mind is that by definition a topological qubit of Majorana, a fermion sort of based qubit, or do you think spin liquids may afford other advantages in the characteristics of the materials, longer coherence times or other means of providing a better foundation for fabricating qubits?
Arnab Banerjee (22:03)
It may, so ⁓ this is basically research that will, some of it will still need to be done. There are actually frameworks for quantum spin liquid based qubits. ⁓ One of them was just published in physical review X by Jason Alicia from Caltech. ⁓ So ⁓ there are frameworks for that. Now ⁓ the engineering towards that will now need to catch up of course.
Sebastian Hassinger (22:15)
Hmm.
Sebastian Hassinger (22:32)
⁓ As is always the case. ⁓
Arnab Banerjee (22:32)
⁓ And ⁓ as is always the case. ⁓ And what I am hoping is that we also need to understand not only the engineering, we also need to have enough confidence on the materials to pour in, let's say, ⁓ a couple of hundred of million dollars on that research so that it develops. And I think that what ⁓ the quantum computing research can help is to provide that level of confidence.
Sebastian Hassinger (22:51)
Right, Yeah. ⁓
Arnab Banerjee (23:00)
that modicum of confidence that actually brings in ⁓ that ⁓ funding into the space to develop ⁓ these new systems. having seen that, Marana qubits, as you know, are being pursued by ⁓ Microsoft ⁓ with varying degree of success ⁓ and ⁓ protected qubits ⁓ are being pursued by everybody, by Continuum, by Google.
Sebastian Hassinger (23:00)
Hmm.
Sebastian Hassinger (23:05)
Right.
Sebastian Hassinger (23:10)
Right.
Sebastian Hassinger (23:16)
Mm-hmm.
Sebastian Hassinger (23:29)
Mm-hmm. Yeah.
Arnab Banerjee (23:31)
All of these are actually some type of a surface code, ⁓ some kind of a topological code which they are employed.
Sebastian Hassinger (23:35)
Hmm. Right. That is fascinating to me that, that, ⁓ surface code is, is just a logical set of operations that are in a sense, simulating a topological state of matter. And, potentially you could achieve a topological state of matter to instantiate a qubit physically and have that protected logic sort of embedded in the, in the design essentially of the, of the material. And then the qubit that's.
Arnab Banerjee (23:51)
Exactly.
Sebastian Hassinger (24:05)
realized in the material. It's so fascinating. I love that feedback loop of ⁓ quantum computers helping with quantum physics experimentation and research, which then helps making better quantum computers. ⁓ When we first met, you were doing a simulation in collaboration with Cuera on the Aquila device. was a Shastri Sutherland material simulation. You've done work with the
Arnab Banerjee (24:06)
Exactly.
Sebastian Hassinger (24:33)
⁓ D wave quantum annealing system as well. Now you're working with IBM system Do you see sort of the the modalities as all having different roles to play at this stage and in the in the experimentation?
Arnab Banerjee (24:47)
This is actually a very important question and ⁓ I have been asked by many people ⁓ and actually I am ⁓ one of the groups who has worked with ⁓ many of these modalities ⁓ and ⁓ each back-end has its ⁓ each modality has its advantages and disadvantages. I'll give you an example ⁓ IBM has qubits which are extremely fast they are transplant based programmable
Sebastian Hassinger (24:58)
Mm-hmm.
Arnab Banerjee (25:16)
and ⁓ but they have a 2D lattice. ⁓ For the research that we are doing, it actually requires many, many quick, cheap qubits. So that's a bit fantastic. That's ⁓ the operation. So that's fantastic for the research. So that's a great fit. ⁓ Something like Continuum, for example, has a much more all-to-all connectivity, but they are slower. ⁓ So if somebody really needs that all-to-all connection for some...
Sebastian Hassinger (25:27)
Mm. Mm.
Arnab Banerjee (25:44)
kind of Hamiltonian, then they will better be using that. On the other hand, if somebody says that, I don't need so much of programmability of ⁓ XXY, I don't need to measure in XYZ, all kinds of bases, ⁓ Heisenberg is not required. I only need an effective model, which is Ising-like and a transverse field Ising is gonna be, make good for me. Then I think that,
Arnab Banerjee (26:13)
D-Wave and Q-ERA will, ⁓ depending on how quantum you need to be ⁓ in the overall scale and how quick you need to be, I think that they offer ⁓ better products. So it really depends on the problem that we are trying to solve ⁓ and ⁓ we really need to fit the ⁓ backend modality to the problem. And this is where I think the success can really happen.
Sebastian Hassinger (26:30)
Hmm.
Sebastian Hassinger (26:36)
Mm-hmm.
Sebastian Hassinger (26:40)
That's really interesting. You've written a roadmap paper on materials for quantum technologies. What do you think the pipeline looks like from your perspective, having done these experiments on a variety of different modalities and being very close to obviously coming from the physical experimentation and now doing simulation? How far are we from
Sebastian Hassinger (27:04)
discovering or improving materials ⁓ for direct use in the design and fabrication of qubits do you think?
Arnab Banerjee (27:12)
⁓ so I think that as far as quantum ⁓ helping is concerned, we are actually quite close. will, I'm actually ⁓ monitoring this space of going into error corrected qubits or error corrected logical qubits. I mean, I am both optimistic and ⁓ cautious at the same time because ⁓ optimistic because it will probably be better qubits, but ⁓ cautious because the effective number of qubits will shrink.
Sebastian Hassinger (27:34)
Of course. ⁓
Arnab Banerjee (27:43)
and the effective ⁓ kind of operations that you can do with those qubits will also probably change. So ⁓ I think that ⁓ the trajectory of this research is going to change over the next two, three years as we go into that space. ⁓ And ⁓ it will be interesting to see where we land at the other side of it. Having said that, ⁓ I think that ⁓
Sebastian Hassinger (27:43)
Right.
Sebastian Hassinger (27:59)
Hmm.
Arnab Banerjee (28:09)
⁓ The goal is to basically make sure that maybe 70 or 80 or 90 qubits ⁓ in some kind of a 2D geometry ⁓ is ⁓ coherently coupled for ⁓ an amount of time, ⁓ which is probably like J by 20, where J ⁓ is the interaction scale. And then we are already getting into a point where
Sebastian Hassinger (28:34)
Hmm.
Arnab Banerjee (28:38)
there are classical techniques that will not be able to do it. So from that standpoint, I think that we are just a factor of do-off, if we need to.
Sebastian Hassinger (28:42)
Mm-hmm.
Sebastian Hassinger (28:48)
Right. And sorry, do you mean 70 to 90 cubits, ⁓ logical cubits that are fully protected with surface code? Or do you mean sort of good enough cubits ⁓ with enough connectivity? Yeah. Yeah. Yeah. So maybe, maybe error mitigation or high enough fidelity for the two cubic gates ⁓ and enough connectivity. Yeah. Yeah. That's, that's not that far off. ⁓ That's interesting. That's really interesting.
Arnab Banerjee (28:58)
good enough cubits good enough cubits
Arnab Banerjee (29:05)
⁓ Correct, correct, ⁓ And from that standpoint, yeah. Exactly. Yeah. ⁓ The choice of problems would be ⁓ important though ⁓ in that space.
Sebastian Hassinger (29:21)
Of course.
Sebastian Hassinger (29:23)
Yeah, I mean, I think that's where your set of experiences becomes really, really crucial because you can draw on your experimental experience and tailor the experiment to the characteristics of that system to choose a problem that is not classically accessible, but is accessible to that particular systems architecture and characteristics. So that's really exciting. ⁓ so, okay, last question you've, you know, with, with the
Sebastian Hassinger (29:51)
years of physical experimentation with neutron scattering. we, you know, forget about 70 to 90 good enough qubits. If you say, okay, you've got a quantum computer of arbitrary quality and scale. What's the material science of the neutron scattering problem you would throw that system at first? What's the answer you want to get overall?
Arnab Banerjee (30:12)
⁓ Well, there are quite so many, ⁓ I think that ⁓ I would ⁓ like to simulate the entire standard model using a ⁓ quantum computer.
Sebastian Hassinger (30:17)
Yeah. ⁓
Arnab Banerjee (30:35)
Get everything. ⁓
Sebastian Hassinger (30:35)
my that's a modest goal. Yeah. ⁓ Yeah, theory of everything. Okay. Yeah, I think you're you're not alone in that. I think that's what everybody but so that would be but that would be from the perspective of starting with material science and building up essentially, right? You would be
Arnab Banerjee (30:54)
⁓ Right, I mean more realistically, I think that there are problems which are completely untractable using classical computers that we can tackle. I mean the frustrated triangular lattice, it's so simple to think about it, but it's so inaccessible. ⁓ And ⁓ any model where you have ⁓ more connectivity, the more the connectivity you have the the quantum. ⁓
Sebastian Hassinger (31:10)
Hmm.
Sebastian Hassinger (31:16)
Right.
Arnab Banerjee (31:20)
entanglement kind of spread so fast that you are completely clueless about what is happening after ⁓ a certain amount of time. ⁓ And any of these problems become quite accessible. Understanding what's the interaction between spins and electrons, ⁓ it's ⁓ mostly ⁓ a problem ⁓ which ⁓ many theorists don't want to touch.
Sebastian Hassinger (31:37)
Hmm.
Arnab Banerjee (31:45)
And I think that ⁓ those will start getting accessible and we will get a much better understanding of what's going on around us.
Sebastian Hassinger (31:45)
Yeah.
Sebastian Hassinger (31:54)
Mm That's great. I mean, I think, like I said, I think your set of experiences makes you the type of person who's uniquely able to navigate this, this, you know, emerging space, you know, as we get closer and closer to being able to tackle classic, you know, in track, classically intractable problems, I think.
Sebastian Hassinger (32:16)
I'll be watching you because I think you're going to be among the first to get to those really interesting results. ⁓ So yet another really nice piece of work. really, I'm always quite impressed with the work you're doing. So thank you so much for joining me today.
Arnab Banerjee (32:30)
Thank you, it was my pleasure.
