Forget Quantum-Enabled Solutions. It’s time to think of Quantum-Enhanced Innovations. And of the second Quantum revolution. Let’s get this cat out of the bag or, shall we say, the proverbial box
What would the first programmable quantum sensor translate into? Recently, researchers at the University of Innsbruck, Austria designed the first programmable quantum sensor and tested it in the laboratory. They applied techniques from quantum information processing to a measurement problem. It is a leap that promises quantum sensors whose precision can inch very close to the limit set by the laws of nature. While it has been perceived that atomic clocks could run even more accurately by exploiting quantum mechanical entanglement, there seems to be avoided when it comes to realizing robust entanglement for such applications. Is this where tailor-made entanglement can be precisely tuned to real-world requirements? What big leap does the research was done by Peter Zoller’s group (at the Institute of Quantum Optics and Quantum Information at the Austrian Academy of Sciences in Innsbruck) give a nudge for? What else can these research initiatives do to leverage the potential of quantum sensors?
Dr. Christian Marciniak, Department of Experimental Physics, University of Innsbruck helps us to unlock some questions around these emerging, but powerful, species called Quantum Sensors.
Tell us something about ‘quantum entanglement as a concept. How best can it be exploited for real-world applications? Any scenarios where it would work well?
Entanglement is a property that quantum systems can exhibit. Mass is also a property of quantum systems (as well as classical). What I can say is that entanglement does not have a classical analog and it is far more elusive in as much that it is only emergent as an observable property in well-controlled environments. Entanglement is, however, ubiquitous (just like mass is), though not every occurrence is what we would call useful.
We cannot know how to “best” exploit it for applications. How can you best exploit friction or gravity? How can we best exploit computers? There are certainly many useful ways, but have we discovered the best and, if so, which is it?
As to possible scenarios, we are currently at the cusp of moving from what we call quantum-enabled technologies (not limited to sensors) to quantum-enhanced technologies.
Can you explain that a little?
The former category is already an enormous industry. Transistors and lasers, RAM, hard drives (HDD, SSD use different types but both quantum phenomena), virtually everything semi-conductor based. Quantum-enhanced technologies do already exist, though much less common. Commercial offerings exist—for example in cryptography. On the sensor front, notably, LIGO and light microscopy are enhanced by quantum correlations, as you will also find mentioned in our paper. For sensors, as we mention, we foresee that using entanglement to push into quantum-enhanced technologies can yield great benefits once the engineering questions of moving from a laboratory to commercial offerings are solved. Once they are, at least conceptually, there is no restriction on where they could be useful. I would imagine that medical imaging and diagnosis would be an obvious point of improvement. They have traditionally been benefactors, just think MRI. Certainly, also more pure metrology applications like in (atomic) clocks, gravimeters, inertia sensors, and navigation.
If you ‘consider’ quantum materials or devices, you are not an early adopter. You are late to the game.
What spawned this work on quantum sensors? What are you most proud of—in terms of its distinctive accomplishment?
Historically, metrological questions and desires have been strong drives for the development of atomic physics. Laser cooling was strongly motivated by spectroscopy and frequency metrology. In turn, what we have learned in atomic physics has been a strong enabler of quantum information processing (quantum computing). Our work now uses an atomic quantum information processor to enhance metrology—full circle. The idea to use entanglement for metrological benefit is not at all new—more than two decades ago people have toyed with that idea. It has been difficult to get it to work reliably, robustly, and in a way that is “worth the effort”, i.e., where overheads do not outweigh the benefits. Likewise, many proposals did not give specific instructions on how to achieve a given benefit. Even though they could perhaps tell you how much you could possibly gain in using a certain resource, they did not tell you how that could be accomplished.
The fact that now we managed to do this, that is outperform any classical strategy even when considering the resource overhead in a real machine with real imperfections and noise—that would be what I consider a great accomplishment.
What next with this research?
The obvious next step is to extend this work to more than one parameter at a time, specifically two parameters whose generators do not commute. In quantum physics, the Heisenberg uncertainty principle tells us that some observables (parameters) cannot be estimated together without changing the measurement outcome. Position and velocity would be one such pair, but there are many pairs. If I ‘do’ want to determine both where something is and how fast it is at the same time, what is the optimal strategy to do so given the restrictions quantum mechanics imposes? There are again very general answers to this question which are however not practical. What is the best ‘practical’, not necessarily best ‘possible’, way to do this? That would be the next thing to figure out.
Can these sensors be enterprise-ready, scalable, and relevant for industrial applications? How soon? Any suggestions? Like Personalised medicine? Robotic replacements for humans in dangerous areas? Metrology?
They can, of course. Nothing forbids them from being made into products and one would hope such a thing would be taken up by industry. However, the time to market for new technologies is often on the order of 3-5 years after development starts. It is worth pointing out that similar technologies, quantum-enabled not quantum-enhanced, are being progressively pushed onto the market or have been there for a long time. If you rely on GPS or other positioning systems, you are using quantum-enabled sensors.
Are Heisenberg Observer bias, superposition, and location-hidden-variable factors still relevant for entanglement?
I am not sure what Heisenberg’s observer bias is, but I assume you mean the fact that the measurement of one quantity disturbs the outcome of other measurements. This is also called (quantum) back-action and is the reason for what I called Heisenberg uncertainty earlier. More specifically, though this technicality is almost assuredly unimportant for your readership, it would be measurement-disturbance relationships which are more practical but less fundamental. Uncertainty relations are fundamental as per current understanding and as such, yes, they are relevant. More than that, the work on quantum-enhanced (not enabled) sensors is largely focused on shaping uncertainties. The uncertainty principle says that the product of uncertainties in two quantities cannot be smaller than something. a * b > c. It makes no restrictions on a and b, only on their product. Classical strategies lead to a = b, we make a << b. We use entanglement to reduce the cost of making b bigger. As long as I don’t care about b, this is a win.
Superposition is relevant, yes. Entanglement makes no restrictions on superposition and, indeed, in most ways to look at it, there is no entanglement without superposition. Which is not to say they are identical.
Local (not location) hidden variable theories are theories that assume quantum mechanics is wrong or incomplete. They try to find models of reality that are deterministic. I am not sure about all of them (since they are many and not my area of expertise), but all I know does not allow for entanglement.
The University of Waterloo is also working on quantum sensors – and for cancer treatment. Are these sensors something similar?
It’s overly simplistic to say that they work on sensors for cancer treatment. Needles are used for cancer treatment. If I make needles, do I treat cancer? The press release on their website says: “could be used for improved cancer treatment”, which is obviously true. Medical imaging and diagnosis are, as mentioned earlier, a frequent benefactor of improved sensor technology because smaller more accurate sensors mean better diagnosis and more targeted treatments. Quantum sensors can help with this in many ways, yes. However, what they have done is develop a very broadband and efficient light detector. What we have done is develop and implement a sensing paradigm—a new way of improving any type of sensor. They made a new car, we opened a car tuning shop. Bring your car, we will make it run more efficiently, faster, and better—in fact, as good as it can be given what you brought in.
For LIGO, who are already using quantum enhancement, we think it might not be worth it because they are already close to how well you could possibly do. Turns out they only needed hammers.
Should the technology industry seriously consider the potential of quantum materials, quantum memory, and quantum devices? Any thoughts?
They have and they have done so for more than two decades. Quantum technology is not new and it’s everywhere from your home to your workplace. Different people have different estimates based on what they assume to be part of quantum technology, but it is a large industry. If you ‘consider’ quantum materials or devices, you are not an early adopter. You are late to the game.
Could you tell us more about ‘tailored’ entangled states? How can this customisation work for tech innovations?
We use the word tailored to distinguish from what has previously been used as the norm. Entanglement, and its generation, come in many flavors. What has been done in the past was to pick your favorite tool, say a hammer, and use your hammer until hammering doesn’t make things better anymore. This was the first quantum revolution where we looked at nature, saw what quantum systems there are, and thought about how we could use those. This gave us the laser, the transistor, MRI, GPS, and smartphones. Over the past two decades or so we have gained more and more control over quantum systems to the point where we started to engineer (or tailor) them to our desires. Can I ‘make’ a quantum system with the properties I want for a certain task? That is the second quantum revolution, which is happening right now. Our approach, the ‘tailored’ entangled state is: Given all the tools I have, not just hammers. How can I use them in an optimal fashion to make what I want? If you want to build a shelf, you might not just want a hammer. Hammer some nails, saw some boards, glue them together, and do some more hammering. Our programmable quantum sensor is a set of instructions you need to build the shelf given the tools you have.
Now, let us say you are making a medical detector that goes into a person. Maybe some capsule that someone has to swallow with some detector inside that then comes out at some point. There is only so much size this thing can have because, well, someone has to swallow it. You are restricted in space. Ideally, you want to make it smaller so it is less unpleasant. Ideally, this does not come at the cost of making the measurement bad, because they have to swallow it for some purpose. Our tailoring means that you can get better results at the same size or, alternatively, the same results at the smaller size. Or some mix.
Is the overhead aspect significant enough if one does this at scale?
The overhead question is important and depends strongly on what you have at hand, how much it can improve in principle, and whether that is worth it. That has to be decided on a case-by-case basis and cannot be answered in a blanket statement. We have looked at two examples that are very relevant in precision metrology and found that the gains outweigh the overhead, so it is worth doing it. For LIGO, who are already using quantum enhancement, we think it might not be worth it because they are already close to how well you could possibly do. Turns out they only needed hammers.
Dr. Christian Marciniak, Department of Experimental Physics, University of Innsbruck
By Pratima Harigunani