Your hosts, Sebastian Hassinger and Kevin Rowney, interview brilliant research scientists, software developers, engineers and others actively exploring the possibilities of our new quantum era. We will cover topics in quantum computing, networking and sensing, focusing on hardware, algorithms and general theory. The show aims for accessibility - neither of us are physicists! - and we'll try to provide context for the terminology and glimpses at the fascinating history of this new field as it evolves in real time.
Sebastian Hassinger 0:02
The New Quantum Era, a podcast by Sebastian Hassinger
Kevin Rowney 0:06
and Kevin Rowney.
Hey, welcome back to the new quantum era. Today we're doing an interview on a really interesting technology called trapped ion qubit systems, these systems they, they they work with individual ions that become trapped inside controlled radiofrequency oscillations for fine manipulation of these individual ions. And they're then stimulated to different excitation states via photons really interesting tech. Another example of these huge advances in engineering in parallel with some deep and substantial theory. So, cool tech.
Sebastian Hassinger 1:04
Yeah, absolutely. Kevin, I'm really excited about this episode. Ion traps are sort of, you know, one of the big two modalities for building qubits, the other being superconducting. We've had a number of episodes where we've talked to people about superconducting qubits. So I'm really looking forward to digging into trapped ions. They're used by a number of private companies, including IonQ, and Quantinuum and they go - their history goes quite a bit quite quite far back. I think NIST was the first lab where a ion trap was used for a qubit. So
Kevin Rowney 1:38
—the original tech wasn't even, you know, oriented towards quantum. But it's physics physics lab. That's right. It was just off the shelf tech, right. Yeah.
Sebastian Hassinger 1:48
That's right. Yeah. And so in this case, we're talking to another public sector researcher, Daniel Stick, who's a researcher at the DOE National Labs, and he and the rest of the team at Sandia National Lab in question. Were recently in the news with some innovations around larger scale traps than they previously been built. With the the project had the amusing name of the enchilada which attracted our attention. So we reached out to Daniel, mostly because we're interested in the technology secondarily because of the name. And so the project was part of the work that's been done at the quantum systems accelerator, which is one of the centers established by the US National Quantum initiative enacted in 2018. The NQIhas driven a ton of really valuable interactions between public and private sector researchers and labs and is about to be renewed. So we're really happy to provide this peek into both trapped ions as qubits and the valuable work that's been contributed by the team at Sandia. So here we go.
Sebastian Hassinger 3:19
Welcome back to the podcast. We're joined today by a very special guest from Sandia National Labs. He's a distinguished member of technical staff. His area of expertise is ion trap research. He has a PhD from University of Michigan, and in 2012, received the Presidential Early Career Award. So quite a distinguished career path. So Daniel Stick, welcome to the podcast. Thanks for joining us.
Daniel Stick 3:52
Thanks so much for having me.
Sebastian Hassinger 3:54
And so, Dan, normally, we like to give our guests a chance to sort of introduce themselves and talk a little bit about your career path, how you got to where you are in quantum technology. So if you want to take us through that, that'd be great.
Daniel Stick 4:07
Yeah, I'd be happy to. So um, let's see, I started in grad school at the University of Michigan, in 2002, in Chris Monroe’s group, he was — it was a relatively new group that he started there. And I wish I could say that I was being really strategic and had a lot of foresight and thinking that quantum computing was going to grow as big as it did. But that was not the case. My undergrad advisor suggested I look into his group and I was really good fit and I'm very fortunate that I kind of started off in that career trajectory. The first project I worked on in Chris's group was to make a microfabricated ion trap and so, this was had to travel to a cleanroom and learn about processing gallium arsenide and made the first monolithic microfabricated I and trap, which was a novel for the time, but it was quickly supplanted by a more useful variety of it called a surface ion trap, that that was just more amenable to microfabrication. And so what I, what I took away from that experience was I just, I really enjoyed working in technology development for quantum computing. And that kind of set me down this path. And it was really important work at the time, because for trapped ion systems, people were making devices out of gold coated ceramics, you could bend wires in a certain direction to make an ion trap, we had an undergrad who sort of twisted a paperclip together and stuck it into an RF source and was able to trap dust particles. So there were just there's a whole host of weird and random things you could do to make an ion trap. But none of them had the prospect for scaling to the larger to the numbers that people were talking about for a quantum computer. And so I it was an important demonstration to start making those things using lithographic techniques and CMOS processing techniques.
Sebastian Hassinger 6:12
So some some limitations to the paperclip ion trap is what I’m hearing
Daniel Stick 6:18
it made me think that I wasn't that special when she kind of bent this thing together and dumped some dust in it. And we all hovered around and saw how the dust floated there in midair.
Sebastian Hassinger 6:28
That's incredible. So I mean, that's, that's really interesting. Chris Monroe, of course, one of the founders of IonQ. So a real pioneer of trapped ion quantum computing, it's it. That gives me definitely more context for the fact that you've been involved in some very interesting research. Most recently, in August, we — this is how we made contact with you, we saw the story about the release of the project at Sandia that's nicknamed the enchilada, which is quite a large scale ion trap, right?
Daniel Stick 7:01
Yeah, that's correct. So kind of continuing that story, the fabrication involved more and more complicated ion traps, from from those first microfabricated versions that came out in, say, 2006. And people have just been making them in more and more sophisticated ways. So one of the key components is to make a junction. So instead of just having a linear chain of ions that you can maybe move up and down in one dimension, people started making versions of that with junctions. And we made one of those at Sandia. And recent demonstrations have shown that you can transport ions to those types of structures at high speeds with low excitation. So that's great. That's an important building block for what people oftentimes envision that trapped ion quantum computer will look like. And so the enchilada trap was an effort to make a device that could just store more ions, move them around in a fashion that would enable more complicated, sophisticated algorithms. And so it contains multiple ones of these junctions. And right now we're working on testing it to be able to do those sorts of movement primitive motions.
Kevin Rowney 8:16
So cool. There's numerous different architectures, right for embodying qubits and doing computation with them. I mean, just for the benefit of our audience, I think it'd be helpful to know I mean, how are your basic functions like gates, entanglement and measurement? How are those realized within the, you know, the ionic systems?
Daniel Stick 8:39
Yeah, so for the most part, that are realized with lasers that are focused in applied to the individual ions. And so in the case of, there are a couple of things that you have to fulfill in order to make a good qubit. They're often called the Divincenzo criteria. And one of them is you have to be able to initialize the qubit into a known state. So we apply a laser, typically, that optically pumps it to a single atomic state, and that's there we go, we have our initialized qubit, then you have to, you'd have to be able to do single qubit operations where you can change the state put it into a superposition. And to do that, we often use lasers, and we apply them for a certain amount of time and a certain amount of intensity at a certain frequency. And by doing so we can perform what's called a single qubit gate. Now other other groups can use microwaves and some people are trying to figure out how to use and have made really impressive progress in applying microwaves to electrical leads on a chip that can then do those same sorts of manipulations and they're working out the challenges with individually addressing. Lasers are easy to focus. Microwaves tend to kind of go everywhere and so that is one of the challenges with applying microwave signals for qubit manipulation to to a small ion trap. Okay, so we've covered initialization and single qubit gates, you also have to be able to do entangling gates. And that's, in some ways, much like the single qubit gate, but it typically, but for one involves two qubits, and you have to use it takes more laser power. And typically they're longer. But at the end of the day, you get an entangled state between the two ions, it instead of just operating on the motional, or on the internal atomic energy levels of the ion, that laser sort of temporarily swaps the quantum information stored in an ion into its motional state. So that two ions in the well. When you apply that laser, the the information is stored in the motional state, and they share that that share that wealth through there, they share that interaction through their Coulomb repulsion, because they're very close together, they're a few microns apart. And then you apply the laser for a particular period of time, and it puts the quantum information back in the internal atomic states of the ion and at the end of the day, you have an entangled state of those two ions. But that that that Coulomb repulsion is the mediating mediating interaction between the ions that allows for entanglement. And then the final thing you have to do is be able to detect which state the ion is in. And you do that by applying a laser that is resonant to a single atomic transition of one of your qubits but not the other. And when that iron is fluorescing, you can see glow on a camera or a photomultiplier tube. And if you get enough counts, then you say, Okay, I know which state it's in.
Sebastian Hassinger 11:41
And you mentioned that first project, in Monroe's group being a monolithic trap. And monolithic, of course, referring back to sort of the birth of integrated circuits in classical computing, it seems like there must be a lot of challenges in trying to combine photonics and RF, you know, in a CMOS setting is that that's sort of the that's a very unique sort of mix of technologies on one chip, right?
Daniel Stick 12:11
Yeah, that is one of the big challenges with ion trap quantum computing, in particular, it requires two, sometimes two very different and sometimes not particularly compatible signals that have to make it onto the chip, optical signals and electronic signals. And so the optical part of that was traditionally done with free space optics. So we would focus laser beams, they would just go across the surface of the chip, with one challenge, the scalability in that scheme is that, as you if you want more ions, the number of ions will kind of grow as the area of your chip. But the number of IO interconnects only will grow as roughly the perimeter length of your chip. And so you've got this problem that quickly becomes challenging. And if you want to focus a laser beam across your chip, you're going to hit multiple ions if they're not in a single line. And so the the way that people have come up to address that and resolve that challenge is to incorporate integrated waveguides onto a chip and route the light to the ion and then deliver the light out of the chip. So it kind of comes out of the plane and doesn't hit another ion. And some of those seminal experiments were done back in 2016, really pioneered by Lincoln, MIT Lincoln Labs. And since then, a number of groups, including Sandia have pushed different aspects of that forward made some advances in lowering the loss of those waveguides. And just demonstrating them within the context of the trap ion.
Kevin Rowney 13:51
I mean, it's an interesting area of engineering. I mean, these magneto-optical systems, it feels like I mean, I'm really an amateur, but it feels like there's major innovations going on in that space. There's a maybe an acceleration of the achievements of engineering in that area. What do you think?
Daniel Stick 14:07
Oh, yeah, absolutely. I think that the development of surface traps certainly kind of sparked this interest in CMOS technologies. But then once people realize that look, this is also a great platform for delivering all these signals around the trap. A lot of other technologies got incorporated. So I mentioned the waveguides groups that various groups have worked on integrating detectors into the traps of superconducting nanowires work that's often done it that was done at NIST, as well as single photon avalanche diodes for detection. And so those are single site elements, you can put them right below an ion. In addition, we've worked here at Sandia on building optical modulators that that are CMOS compatible and can be built directly with the trap. And then some other groups are working on I'm building electronics right in the trap. So kind of coming full circle and taking this device which was had no active electronics on it. So that's maybe a common misconception is that the ion traps that were first built as surface traps, at least at Sandia only employed what's called the back end of line in terms of CMOS processing. That means it's just metal oxide, metal oxide, it did not include any active or doped regions. And so, you know, we were not using very much of the CMOS capability. But as time has gone on, people have thought, well, now Now let's put other No, that's dope region and make it a single photon avalanche diode, or make some sort of active DAC built directly into the trap that can control some electrodes.
Sebastian Hassinger 15:47
That's very cool. And the the work itself so that the enchilada trap, I think, is part of the Quantum Systems Accelerator at the NQI says the National Quantum Initiative center that you are part of as a collaboration with Duke and others, I believe, is that right?
Daniel Stick 16:04
Yeah, that's correct. The quantum system accelerator includes researchers, at many institutions, I won't start to name them lest I omit one of them but but there are researchers who were work on neutral atom quantum computing, researchers who work on superconducting quantum computing. And then there's the ion trapping groups. And there are multiple groups involved at NIST at Berkeley at Duke at Sandia at MIT Lincoln Labs. Jeez, I hope I don't forget somebody, but
Sebastian Hassinger 16:38
NQI centers always have a webpage. It's like all the participating institutions. It's like a NASCAR sponsorship page, right?
Daniel Stick 16:46
Checking the web page, don't hound me if I forgot
Kevin Rowney 16:49
Trying hard. Yeah. Yeah.
Sebastian Hassinger 16:52
I love the fact that the the NPI centers are so focused on that, that broad cross institutional kind of collaboration that as you said, I mean, every everybody is sort of bringing their own strengths, and has their own perspective and brought to bear. So the the enchilada, how many qubits have you? Or is it capable of sort of realizing that architecture?
Daniel Stick 17:17
Yeah. So I guess first of all, say something that maybe answer a question that begs to be asked is the the name enchilada was not chosen for marketing reasons, although it's turned out to be pretty popular here in New Mexico. And it's just because we we design these traps, we spend several months designing them, somebody lays it out somebody, people process them for nine to 12 months, and then we have to put them in a package. And we're doing this for several traps at the same time. And so instead of remembering that it's called RS1973, we make up a name that we can refer to it. So that's where the name enchilada came from,
Sebastian Hassinger 18:01
and someone was hungry, when it came time to name this.
Daniel Stick 18:05
There are going to be two versions of it. And so I think I had in my mind that on the second version, we would pile all of the different technologies we couldn't into it and make it the whole enchilada. So that being said, it was designed in collaboration with the Duke group, they have a particular architecture in their mind for what a trap buying quantum computer looks like. And it involves long strings of ions. So somewhere around 32 ions we’ll say, right, deliver laser beams through a multi channel acoustic optical modulator that can hit any of the ions in that string, and do gates between any of those ions and that strength, that's a really powerful tool, because they have what they call all the all conductivity. They don't have to move the ion on the right side of the string over the ion and the left side of the string in order to interact them, they can just apply laser beams to those two ions — just is not the right word, they have to do a lot of calculation of how the laser pulses look like in order to achieve high fidelities. But that is kind of their mode of operation.
Kevin Rowney 19:13
But that seems like that's a huge advantage relative to other quantum computing architectures. Yeah. I mean, in many of these other quantum computer architectures, I think they go to pretty operatic lengths of engineering to promote interaction between far distant qubits that are separate in the physical location on the chip. Do I have that right?
Daniel Stick 19:33
Yeah, it's definitely a strength. Not only is it a time saver, but it also kind of saves on fidelity because instead of having to do 28 operations, where you may have some error for each operation all of that has to be kind of combined in the context of like, how does one run a real algorithm with this and if we want to go from 32 ions to 64 to a thousand someday, what are all the different features that have to be developed and successfully operated in order to make that work. And so it is a very powerful thing to be able to do those have that all kind of connectivity within a string of ions. But there will come a time where they have to kind of figured out how to connect those strings together. And one of their ideas that they've promoted and others have done a lot of work on is the idea of using photonic interconnects, where they collect photons from a single ion collect photons from a distant ion, and then they interfere those in order to project the two ions in into into an entangled state. And so that's a really also a powerful tool. But we think that you will also still out there still some room for the traditional shuttling of ions around the track in order to get them to interact with ions that are not necessarily within their chain, or nearest neighbors. And so the enchilada trap has, you know, kind of has this looks like a TIE Fighter kind of, and if you had chains on each of the four corners, or the four legs at the extremities, and you, you could divide them and move parts of those chains into this middle interaction region. You could combine chains in ways that allow you to do gates and run out rhythms on the collective group of ions rather than just a single chain.
Sebastian Hassinger 21:26
So each each one dimensional trap, is there still limitation around 30, 32 ions in a single trap? Because of thermodynamics, I think, right? Or is it some other? Yeah,
Daniel Stick 21:37
yeah, yeah, there are some limitations, the heating rates get worse when you increase the number of ions. At some point you, I guess I don't know where that is. But making acoustal optic modulators that are longer and longer and longer becomes challenging. So there is some natural and technology limited sides of those chains.
Sebastian Hassinger 21:58
So in that shuttling or you're bringing individual ions from two different traps, two distinct traps to into some sort of proximity, and then in doing the entangle, how's the entanglement happen from one track to another?
Daniel Stick 22:13
Well, once you bring them into the same region, and you apply a laser pulse, which is hitting the two ions you want to entangle, they will be entangled. And once they move apart, they will remain in Tangled provide, they see, we took care of a lot of other things. And so that's that is one of the reasons for making a trap like this, because we don't know necessarily what are the problems that we are introducing by making a larger trap. So for example, as the traps become larger spatially, they will be sensitive to magnetic field variations over space. And so simply moving the ions through those magnetic fields could have an impact on the quantum information within the ions. Hopefully, that's not a big problem. If we use certain atomic states that are not magnetically sensitive, that will not be a big challenge. But you know, it's you never know what's going to happen with this.
Kevin Rowney 23:07
This is great detail. Thank you so much. And just so we make it clear for our audience, these are these are not devices that run deep within a dilution fridge. Yes. I mean, they're all of that all that infrastructure is not needed to this architecture. Yes,
Daniel Stick 23:22
that's correct. That's, that is one of the selling points for trapped by in quantum computing is that there is no threshold temperature at which you make the qubit go from behaving really well to behaving, you know, above which things would operate really poorly. The advantage? Yeah, decoherence. Yeah, the advantage of operating ion traps in cryogenic environments is it reduces this air source called anomalous heating, which you've talked about with previous guests on your podcast, it also increases the just the physical lifetime of the ions because it reduces the background, gas pressure limits the number of collisions that occur. And so we've had, we've had a crash that in our, in our lab where we trapped the single ion and we're able to hold it for, I think, six days in an ion trap. Normally, it's not that that good in that room temperature trap will hold an ion for six hours or something.
Sebastian Hassinger 24:23
Yeah, how long is that an ion years?
Daniel Stick 24:27
Depends on how fast the gates are.
Kevin Rowney 24:30
Right? So this is great. And this is sorry. So there's all these cool trade offs that emerged in different architectures. It feels like there's some distinct strengths in this architecture. I mean, certainly in terms of coherence time, and certainly in terms of disposal of lives of excess infrastructure, like a fridge, but what are their Are there limits in terms of like, oral comparisons where it's not so favorable in terms of speed or accuracy or other aspects of how you would evaluate a QC system?
Daniel Stick 24:59
Yeah, so If this the enchilada device itself can support other other modes of operation for trap buying quantum computers, so the other kind of competing version, or I don't know, they're kind of on a spectrum, but the other version is called a quantum CCD. And it relies, instead of having long chains of ions with the prospect for having photonic interconnects, it requires really on single ions that just move to their nearest neighbors. And moving it in a certain pattern, you can either implement an algorithm that is customized for that device, or you can implement something like quantum error correction. And so that's that the the end of the spectrum that requires a lot of ion movement in order to work, the the version I talked about with the photonic interconnects would leverage other things like these, these both entangled photons to make some of those connections. But the idea is that one of the ideas that the enchilada trap can, at this, at this scale accommodate both versions of that. Cool.
Sebastian Hassinger 26:03
And so in the press release the Sandia press release, there is a mention of up to 200 qubits. So if you've got four, three, roughly 30, qubit, traps in in that in that TIE Fighter array, as you said, on the on the monolithic chip, is there an is there an assumption there's an interconnect between that and another chip that has another four traps on it,
Daniel Stick 26:27
yet? No, the assumption is that the other 18 Missing ions are doing something else that is actually also useful. So for instance, sympathetically cooling, what that refers to is that if you leave an ion sitting there in space, and you don't actively Cool it, it's it will heat up over time due to this problem of an anomalous heating. And then whatever, entangling gates you play after it, after that point, will have very low fidelity because it relies on swapping information between the internal states and emotional modes. And if the motional mode is really hot, you're going to have a low fidelity process. So one might think, Well, why don't we just cool those ions, like we do normally, with Doppler cooling or something? Well, that that causes decoherence, that immediately destroys the quantum information of that ions. So we cannot apply a Doppler cooling laser, to the ions that we're relying on distort quantum information. And one of the ways people have figured out to get around that is to what's called sympathetically cool, and so that means they store other ions in the same well as the ions that they want to keep cold, and they cool those other ions. And those other ions don't care about the information that's in there. All they're doing is sucking up motional energy that is accumulating within the chain. Interesting. So that's where you get up to larger and larger numbers of ions.
Kevin Rowney 27:56
And so much of this is based upon a CMOS manufacturing techniques. To me, it feels like there's a lot of powerful computational possibilities, done a route that is, you know, really already equipped to provide almost consumer grade access. I mean, is it? Is this a route towards someday being perhaps a desktop system for for home applications either want to talk to? I mean, I don't know. It feels like it's, there's some room for optimism. No.
Daniel Stick 28:24
Yeah. I mean, I'm not famous enough to be the person who was the IBM person who said something like, there might be 50 of these. Right. But nonetheless, I don't think I will — I guess I'm would be surprised if this reached the individual user market. But I guess I'll say a couple of things about that. One is the and I'm not sure any of these were, people realized how, how unique regular classical electronics were. But for quantum computing, you really need a multiple analog signals per ion, or qubit, whatever you have. And so to generate those analog signals, with the precision and speed that you need in order to run a powerful computation, that is just an enormous amount of other electronic information.
Kevin Rowney 29:24
Yeah. And so that none of it cheap, none of it easy to maintain.
Daniel Stick 29:29
None of it cheap, these lasers are expensive, they took a long time to maintain, or they take a lot of effort to maintain the even just the RF control signals that we use to modulate those laser beams. Those are not a small footprint. So if we're talking you know, people have rack mounted quantum computers and they're starting to develop that the these companies, but if you then multiply that by a factor of 100, it's not totally clear to me how you're going to maintain performance while limiting the size to something that would fit in somebody's house. Yes, yes. Right. And 100 is not the right value. It's multiply that by 10,000.
Kevin Rowney 30:08
Maybe we're talking right? Yeah, that's right.
Daniel Stick 30:13
One of the thing that you brought up is how this is kind of a unique, CMOS application. So a challenge that we had to initially face when making these surface ion traps in Sandia CMOS boundary was, was related to the fact that sea moss electronics uses relatively low voltages, you know, five to 10 volts, typically, and we're talking about operating ion traps in which we have to apply a 200 to 300 volt RF signal to it that said, around 50 megahertz, just in order to store the ion. And so a lot of our fundamental research at the beginning that continues to this day really was revolving around, how do we make these little devices where we have a few microns of oxide that are separating this one electrode that is at 200 volts, and this other electrode that's grounded, without causing breakdown, that causing excessive power dissipation, we had to kind of invent some of those techniques to make these devices suitable for that very unusual application.
Sebastian Hassinger 31:18
And the those that research when you said before you were that the next version is for the ultimate version, this will be the whole enchilada, you were thinking of calling the whole is that sort of incorporating other capabilities by leveraging this research into these, you know, sort of unique applications of CMOS to bring other as you said, electronics or other capabilities onto the onto the trap itself.
Daniel Stick 31:44
Yeah, yeah. So this, this first version, whatever that's called the little enchilada are something that we've really focused on figuring out how can we make this device bigger than our past devices, while still not dissipating excessive amounts of RF power. And so we incorporated a couple of techniques, we have a version where we've perforated the electrodes in order to reduce the capacitance of the device. But those are the things we're going to test with with this current version. The next one, the whole enchilada. One is we're going to route out all of the electrodes that are on the device, there are over 300 electrodes on the device, of which right now we tie a lot of them together, so it fits on a normal package, the second version is going to have all those be independent, so that you can move ions within the trap entirely independently. And then we also want to incorporate in a CMOS compatible way, the photonics that I was talking to you about before like waveguides, maybe some optical modulators. Perhaps some detectors, although those need more development, I think in order for them to be viable, at least the version to make. And so yeah, put it putting those different tools onto the device in order to make it easier to actually address those 200 ions, or however many are stored, will will be the goal of that device.
Kevin Rowney 33:09
So cool. I get the impression that we're actually living through a pretty gigantic era of innovation in CMOS fab techniques for alternate computing platforms. I mean, the whole spintronics area of research, for instance, a primary example of that, but what what other CMOS techniques, are you guys exploring, besides the high voltage transitions? Are there other other new frontiers, you guys are exploring?
Daniel Stick 33:33
Yeah, so a lot of the a lot of the attention right now is focused on optics, just because of the number of optical signals that have to be routed in on these devices. I mentioned these optical modulators. So that's actually kind of almost a MEMS type device where we apply voltages to an aluminum nitride across an aluminum nitride layer that is piezo activated, and it causes these waveguides to kind of potato chip up. And so that drains the waveguides and causes there to be a phase shift and we can use those to make mach zehnder interferometers that control the amplitude and phase of the light going to an ion and so that all
Kevin Rowney 34:14
the way down to the micron level of scale. Yeah, that's amazing. Yeah,
Daniel Stick 34:20
I wish there were micron size, they're more like 100 by a
Kevin Rowney 34:25
forgiven my ignorance, but still, that's right. I mean,
Daniel Stick 34:28
they are they are straining these little these thin little waveguides at levels that that that strain induces a phase shift that work so that's something we're considering. You know, when one common CMOS capability that people talk about a lot is incorporating digital to analog converters so that you can send a digital signal into your vacuum chamber to the chip and have it produce analog voltage outputs. That is, I think that's certainly on the frontiers. And people have done some demonstrations with something like that. But it's pretty hard right now to replicate the performance of what you can buy from, you know, National Instruments or, or somebody else or make out of your own FPGAs. And
Kevin Rowney 35:21
there's new innovations on the horizon that
Daniel Stick 35:25
that is the grand vision is that you've got this chip sitting inside of a chamber, a bunch of a small number of digital signals go into it. And those digital signals, tell your modulators how to modulate the light on each of the channels. They tell the electrodes how to shuttle an ion from one point to another point. They tell the detect— they read out detector signals that correspond to individual photons hitting the detectors. That's the dream. So the whole thing is a
Sebastian Hassinger 35:53
digital package as far as the outside world is concerned,
Kevin Rowney 35:56
sort of like a system on a chip style. Yeah, sure. Yes, that
Daniel Stick 36:00
would be Yeah, it would be very sophisticated system on a chip very.
Sebastian Hassinger 36:03
And I wonder, does that that level of sort of integration of other functions? Does that give the platform the enchilada architecture more flexibility for other applications as well?
Daniel Stick 36:17
Yeah, I that's one of the I think they're really redeeming things about trapped by in quantum computing is that there are other there are other applications that that can employ and develop the same technologies and advances that trapped on quantum computing needs. So one of the other projects we have here at Sandia, led by my colleague, Hayden McGinnis, is to make a multi ion clock in which you have a whole bunch of ions that are stored on a chip, and you separately interrogate those ions with a laser that you then discipline and lock into that optical that that laser becomes your optical source, and forms a clock. And so this is, this was a kind of a theoretical technique that some people described. And it differs from the traditional optical clock in which the idea was to just put more and more atoms that you interrogate, with a single laser, or really take advantage of the kind of individual addressing you can get from some of these systems. So that you can split the light up, you can, you can interrogate for different times at different parts of the chip, you can interrogate in and out of phase, so that you detect some ions while you're interrogating other ions. And by combining all these techniques, you can make a more accurate atomic clock
Sebastian Hassinger 37:39
even more accurate!
Kevin Rowney 37:42
really anticipate, like more breakthroughs on accuracy. I mean, there was very recently a pretty big upgrade, right in the accuracy of atomic clocks. Is there more frontier to cover?
Daniel Stick 37:52
I think there are. But I think that there are also other dimensions of performance improvements you can imagine. So I think it's more likely that this kind of multi ensemble scheme will help make more deployable, smaller clocks that don't actually have the same performance as say that 10 to the 19 accuracy clock that they're running at JILA at NIST, but maybe it gets the 10 to the minus 16 accuracy, but it can be made small and liable and manufacturable, because it leverages some of the things that are built onto the chip.
Sebastian Hassinger 38:27
What's a few exponents between friends anyway? So yeah,
Daniel Stick 38:31
we wouldn't want to steal the record for from them.
Sebastian Hassinger 38:35
I think for most users, though, ten to the 16 is probably accurate.
Kevin Rowney 38:40
Yeah, it's like plus or minus one second for the entire lifetime of the sun. So pretty good. Yeah.
Daniel Stick 38:47
There'll be happy with that. Yeah, not bad.
Sebastian Hassinger 38:49
Yeah. And what about communications and networking?
Daniel Stick 38:54
Yeah, so I mean, that is a promising application for trapped ions, because they they have their wonderful long term memory that could be used at a at a node that's used to connect the larger quantum network, most of the proposals and things that I think about involve using that those nodes more as things that connect to an individual quantum computer into a larger and more powerful quantum computer, but they certainly have have the prospects for the communication side of things, right? Yes.
Kevin Rowney 39:26
So you mean like long distance secure quantum communication? Yes, right. It's
Daniel Stick 39:31
one of the huge challenges with trapped ions is that it's very inefficient to collect the photons from an ion that is emitting them. So you can put a big gigantic optic above your ion. And you might collect 10 to 20% of the photons that are coming out at a prescribed rate based on how quickly you excite that ion. And 10 to 20% even as I'm saying now seems like a pretty read suitably efficient number. But when you add all the other inefficiencies and losses and fibers and things like that, the rate at which you can establish remote entanglement between ions typically is a little bit slower than you would want. So sub kilohertz, this could be an area actually, though, where CMOS fabrication of many ion traps helps that problem by accepting that you get one kilohertz type rates from a single ion pair here and 10 kilometers away, but then multiplying that by 1000, because you put 1000 of these ions on the same chip, that you're all trying to operate in sequence, you know, make it more capable by expanding the number of ions that are trying to do this.
Sebastian Hassinger 40:48
Is it a similar sort of, you know, calculations, a trade off in your mind, we talked a little bit about superconducting obviously requires a dil fridge and has more limited connectivity compared to the old all in an ion trap, but has much faster clock speed or gate times, essentially, is that sort of how you think about it in your mind as well, just as you described with networking, you can compensate for the this system speed by by by having more capacity, essentially and doing parallel?
Daniel Stick 41:26
Yeah, yeah, that is. Yeah, that is one of the things that we have to think about is our gates are just not 100 times to 1000 times slower than superconducting quantum computing systems or solid state quantum computing systems, and ways to get around that have to leverage other kinds of other attempts that are not limited by the physical speeds that are possible with with an ion trap that we're, you know, that are not limited just by the physical nature of the mind. And so yeah, scaling the systems just by duplicating them and replicating them. And using all of those things in parallel, is one way to do that.
Sebastian Hassinger 42:06
It's very interesting. And so what's the next sort of phase? In your mind? What's the next goal? The do? Are you shooting for the whole enchilada? Are you shifting some other type of Mexican dishes? What?
Daniel Stick 42:24
If I do, I'll think about the name more carefully before I —
You know, we've Sandia has had to, I guess, think I at least I have had to think more strategically about what, what are the useful things we can do as a national lab in a time period in which there are companies that are making amazing advances for system developing systems. And, you know, I'm very happy that there are those companies out there, because they are really pushing the boundaries of robust systems, they have to have high uptime, if they want to offer a cloud surface service or accommodate many users. They can't be like our lab, which is, you know, it operates for a few hours a day, and then we've got to tweak things, they're really ironing out a lot of the difficulties with right, long uptime. And so the, where was I going with that? Oh, so So I, I've had to think what what can Sandia do that is really useful in this context. So my conclusion is that we can investigate different technologies, Integrated Technologies, like I've talked about with the modulators that we're not going to make 10,000 of them, we're not going to compete with companies that are trying to make an entire product, but we're going to test these individual components, sometimes you're going to combine multiple, multiple ones of them together to see whether they behave well together, whether they can be co fabricated, and, in that way, kind of push the technology forward. And we can also try things that, you know, if if, if a company is like has decided to go down one path, they they've made a lot of trade offs and decisions at that point, that are, you know, that they've had to make, but that could be limiting. We can try different things, maybe that could ultimately be useful to those companies in the future or could be useful to the broader research community in the United States.
Sebastian Hassinger 44:30
That makes so much sense to me. I mean, it feels like that's the optimal Public Private Partnership kind of configuration where the public sector is still exploring all those potential, you know, high risk, high reward kind of ideas at the foundational roots of the technology. And private industry is sort of trying to productize or ruggedized or scale the technologies and move them up that the TRL scale right to something that's that It's a mass marketable?
Daniel Stick 45:02
Yeah. Yeah. I mean, I, I'm jealous of how robust their systems are, hopefully, some of these technologies, you know, oftentimes when we make these devices, we take like five steps backwards in terms of performance. So like, for instance, the modulators that we demonstrated, they had very similar performance to the acoustal optical modulators that we replaced in terms of the consistency of the pulses, the optical pulses, but they had like 80 dB less extinction, or maybe 60 dB less extinction. And so, you know, we expect that we can do better on that, and we can improve them in the next round. And so we plan to do that. But for all of these things, we're not like, you know, making just a total leap in all directions, forward in terms of the performance of the devices, we hope we can do that with, with regard to the robustness as we as we integrate more things on chip, and we fix their initial problems, then we don't have to align laser beams to the chip, or we don't have to, we're not susceptible to the lab temperature heating up and moving around the opto mechanics that before we were surrounded here, vacuum chamber with. And so I hope some of those developments can also improve uptime and robustness that
Sebastian Hassinger 46:22
I also wonder at some point, if if the, the supply chain vendors of things like lasers, who are benefiting from the increased investment in the private sector may end up with components that you can leverage in the future for future experiments, sort of like either lower cost or smaller, more miniaturized, etc. I mean, because of the level of investment from the private sector, you can sort of share some of the benefits and the advancements in the supply chain, potentially.
Daniel Stick 46:50
Yeah, absolutely. We really, we really hope that's the case. ions, you know, they are slow. So that's a bad thing. But they fit really well, actually, with developments in FPGAs. And systems on chip, they, if you look at kind of the world of homebrew ion trap control systems, there's a lot of activity going on there, because physicists and their friends can work on these FPGAs and RF systems on chip in order to make something that has a really powerful control system, just because, you know, I think the communication rates and timing and accuracies match well with the gate times of a trap ion quantum computer. And so you know, that that is an area where some other giant industry really pushed the performance devices,
Sebastian Hassinger 47:42
5g in particular. 5g is I mean, I've been told that FPGAs are basically available at the price point because of 5g developments.
Daniel Stick 47:52
Because of 5g we have access to this hardware that is really useful for experiments. And hopefully, because a trap and quantum computing company succeeds really well, the cost of lasers at 493 nanometers will come down and we can buy them cheaply. And in quantity.
Kevin Rowney 48:16
I hope so I'm a little bit of a hobbyist amateur science guy. I do experiments here. I mean, is it ever you think possible within the short term timeframe that a home DIY person could take on building?
Sebastian Hassinger 48:30
Do you have a paperclip? Kevin?
Kevin Rowney 48:33
Exactly. Right. Well, I
Daniel Stick 48:35
don't you know, I don't know that the diode lasers have been become have have become much less expensive. I mean, how much money does this home DIY hobbyist have to spend on?
Kevin Rowney 48:47
Hypothetically? Yeah, exactly. Yeah,
Daniel Stick 48:50
yeah. Some things DFB. Lasers, for instance, are relatively low cost, they have narrow alignment that's suitable for ion trapping. And the the electronics are not that expensive. And so I could imagine a time where the lasers become low enough cost that you can imagine building training setups at universities that are not that expensive, that allow people to trap single lions and do an experiment with it.
Kevin Rowney 49:18
Cool. Wow. It's such amazing times we live in many ways. Yeah,
Daniel Stick 49:24
yeah. Yeah. No, I'm, I feel fortunate to have gotten involved in this one. I
Sebastian Hassinger 49:28
did. Good choice. Excellent. Well, thank you so much, Dan. I think that's a great note to end on. I really hope that you're continue on this work. It seems it seems like you're fulfilling the Quantum Systems Accelerator mission to the tee. But it's really exciting to hear about all the potential that this platform brings with it. And thank you so much for explaining to us and thanks for your time.
Kevin Rowney 49:54
We really had a great time. Thank you so much.
Daniel Stick 49:56
Thank you for having me.
Sebastian Hassinger 50:44
Well, that was yet another highly enjoyable, highly illuminating conversation with our guests, I thought that was really, really interesting, in particular was great to start to dig into ion traps as a topic. They're a very popular modality. They're one that have been around since as long as superconducting qubits or longer, I think you could argue. And it was just really interesting to dig into how they work, how they, you know, sort of the state of the art, at least in the, on the research side of things. And also just, it was really striking to me, Kevin, that, you know, that Dan's first project with Chris Monroe was to build a monolithic trap, monolithic, meaning all, you know, integrating on one piece of a when component or one integrated device. And that monolithic solution in the classical computing context is really the moment that heralds the the explosion of the digital revolution
Kevin Rowney 51:44
Dawn of the computing era. Yeah, that's just a microprocessor.
Sebastian Hassinger 51:47
Exactly integrated circuits and microprocessor came out of TI and Fairchild, and really, you know, set the context for all the digital revolution. So it's really interesting, that's very well started to change. Absolutely. So it was so interesting to hear the progress that they're making towards very capable, monolithic systems in the in the ion trap world.
Kevin Rowney 52:10
That was really cool. I thought, I guess another theme that we are seeing it a couple of interviews, saw it very powerfully here is all of the parallel lines of research and the collaboration at a very high level between, you know, world class scientists, and world class engineers. I mean, you have, for instance, all this huge deep level of expertise in CMOS engineering, you know, then as a fellow traveler with, you know, qubit, engineering, I mean, that, that kind of echoing back and forth of those two, powerful and deep levels of, of engineering and research, reinforce each other, propelling them forward. really inspiring. We really live in fascinating times, with that same theme came up a couple of episodes ago on, you know, the parallels also between how, you know, Quantum Information Sciences and machine learning are reinforced right out there. So, really, really rich, fun topic.
Sebastian Hassinger 53:03
Absolutely. And I think, you know, not only those collaborations, but also across public and private lines as well. Really powerful. Yeah, there's still there's such a, an enormous role for the public sector in in this research era, and, and also computing as a discipline as a parallel with sensing and networking and other other quantum applications. Those all those technologies are, are, you know, aiding each other along the way, it seems. Yeah. Yeah.
Kevin Rowney 53:35
Just amazing. Well, I had a good time. It was just fun. It was fun.
Sebastian Hassinger 53:40
Me too. Yeah. So we hope our listeners enjoyed that as well. And if you did, please rate and review wherever you get your podcasts and subscribe so you can catch up on all the episodes that we're planning. We've got a bunch of very exciting guests planned for the future. So we hope you'll join us then.
Kevin Rowney 54:02
Okay, that's it for this episode of The New quantum era, a podcast by Sebastian Hassinger. And Kevin Roni, our cool theme music was composed and play by Omar Costa Hamido. production work is done by our wonderful team over at Podfly. If you're at all like us and enjoy this rich, deep and interesting topic, please subscribe to our podcast on whichever platform you may stream from. And even consider if you'd like what you've heard today, reviewing us on iTunes and or mentioning us on your preferred social media platforms. We're just trying to get the word out on this fascinating topic and we'd really appreciate your help spreading the word building community. Thank you so much for your time.