ZuriQ: Shaping the Future of Penning Ion Trap Quantum Computers

Quantum computing promises to solve complex problems beyond the reach of classical computers, with trapped ions being a leading candidate for robust and high-fidelity qubits. 

While quantum computers with a few trapped ions have already been demonstrated to work amazingly well, the challenge remains to demonstrate the same for quantum computers with orders of magnitude more qubits. 

ZuriQ pioneers a new approach to trapping ions, aiming to arrange tens of thousands in a three-dimensional array and use them as high-quality qubits to show utility in solving commercial problems. The spinout from ETH Zurich was founded by Pavel Hrmo, Tobias Sägesser and Shreyans Jain in the spring of 2024 and is now part of the QAI Ventures accelerator’s cohort II.

Learn more about the future of Penning ion trap quantum computers from our interview with the co-founder and CEO, Pavel Hrmo: 

Why Did You Start ZuriQ?

It’s really exciting to see something that I worked on as a postdoc in academia perform at a level beyond what we first imagined. When you apply for research grants, you hope things go well enough so you’ll get more grant funding to continue your research and improve next time. Instead, our experiment worked out of the box!

The research project focused on how to trap ions in a better, more scalable way so that they can eventually be used as qubits to perform quantum computations. We were so impressed with the outcomes that we decided to use it as the basis for building scalable quantum computers. 

When I started as a researcher, I didn’t plan on founding a startup, but I realized it made much more sense if we wanted to turn our lab demonstrator into an actual, large-scale quantum computer: research labs operate at too small a scale, with just a few PhD students, and thus, technology development would be much too slow for us to afford to build large-scale quantum computers in the face of international competition.

Also, it will require a level of engineering that you can’t do in academia. You need to hire many specialized engineers, not just quantum and optical but also software engineers. A research group has many domain experts but not enough specialized engineers to take care of all the rest, and it typically can’t put it all together and work on one huge project. Founding a company was a logical step to bring all the required expertise and talent together to work on one overall goal: building a large-scale quantum computer. 

Finally, it’s also strange that Switzerland hasn’t seen a major quantum hardware startup so far, given its strong heritage in quantum physics research. They have a huge talent pool with nowhere to go in Switzerland if they want to work on quantum computers, so we’re in an interesting position: we face strong competition from other quantum startups internationally, but we’re one of the first to build the quantum ecosystem here in Switzerland. 

How Does Ion Trap Quantum Computing Work?

We use electromagnetic fields to trap ions, allowing us to encode qubits using their electronic states by selecting two energy levels representing the zero and one quantum states. We then use lasers to manipulate and measure these energy states and execute quantum operations. Trapping the ions allows us to keep them in place and address them with lasers but also ensures they’re well-isolated from the environment and maintain their quantum states for a longer time, making them reliable for quantum computing.

Trapped ions are excellent qubits, providing excellent detection and readout of quantum information, the highest two-qubit gate fidelity in the industry, and large quantum volumes. The real challenge is how to get many qubits in a single trap chip: we have demonstrated a research prototype in the lab with a single ion, which was relatively straightforward and great for testing small-scale operations. Next, our goal is to increase the number of qubits massively.

Traditionally, ions are trapped using radio-frequency fields, known as Paul traps, which confine ions in one-dimensional chains. Our approach is to replace the radiofrequency field with a homogenous magnetic field and a DC quadrupole electric field, a so-called Penning trap. This allows two—and three-dimensional arrangements of ions and much greater flexibility to exchange quantum information amongst them. 

Since ions are charged, they all interact with each other through the Coulomb force and can thereby exchange quantum information. Other quantum hardware platforms, like superconducting qubits, are limited as only qubits right next to each other can exchange quantum information, restricting the kinds of quantum algorithms you can run. Even with more qubits, it becomes increasingly harder to run larger quantum algorithms. As we scale to larger quantum computers, our approach ensures we are still flexible in rearranging them and executing two-qubit gates between them—without sacrificing all the good properties like all-to-all connectivity that made ion traps a successful quantum hardware platform. 

To trap ions, we place electrodes on a chip to apply static electric fields and then place the whole chip in a homogenous magnetic field that is applied parallel to the plane of the electrodes. The ions are trapped where the combined effect of the electric and magnetic fields results in a stable potential well, and the electric field becomes null. So, we get an array of wells, and by engineering the electric fields, we can move the positions of the wells anywhere in three dimensions. 

This works really well, as we’ve demonstrated in a Nature publication. We moved an ion over a large area of 75×40 microns to trace out the ETH logo, showing that we could position the ion accurately and without micromotion.

To perform a two-qubit gate between any two qubits, we can thus move them as close as possible to enhance the Coulomb interaction between them, allowing for better control over the interaction, reducing gate execution times and decoherence. The gates are then performed by illuminating the ions with laser light while they stay in their respective potential wells. 

Our goal is to build a module with an order of magnitude of 10,000 qubits with very high fidelity, which will already allow us to address the first applications, e.g., in drug design or search for better catalysts. Even if you’d manage to squeeze one logical, error-free qubit out of every 50 physics, error-prone qubits, it would take millions of qubits and will be extremely challenging to build error-corrected quantum computers in the next decade. 

We think we can get utility from quantum computers by taking a much more pragmatic approach, trying to squeeze as much computing power from a module as possible. This requires already a lot of hard engineering to optimize the density of qubits or to address all of them with lasers. If we get the design right at the module level, we can then interconnect those modules to build truly large-scale quantum computers in the future.

Using a bigger magnet to place more modules in the same homogeneous magnetic field and then using interconnects between modules will be the way to go. Also, while we’re currently working with sapphire or glass substrates, we want to move on to silicon chips in the future, as they’re easier to manufacture, and we can leverage a lot of existing fab infrastructure. 

With the right kind of laser setup, I’m pretty confident we can demonstrate great fidelity, not just in the research lab where close to four nines, 99.99% of fidelity has already been demonstrated, but more importantly, also for commercial-grade, large-scale quantum computers.

Our core objective for now will be to scale the number of ions we can handle from double to triple digits, even before turning them into qubits. Showing that we can handle thousands of ions in a reconfigurable and manipulable grid will be an important milestone. We will then turn them into qubits and build a quantum module with excellent fidelity. 

How Did You Evaluate Your Startup Idea?

There are two aspects to this: On the one hand, we have a strong R&D roadmap, with clear technical milestones to develop larger-scale quantum computers and crucial IP. This will systematically de-risk our approach, keep investors interested, and help with successive capital raises. 

On the other hand, we complement this roadmap with R&D and government grants and can sell quantum computers for on-premise installation. We can’t exactly plan if and when we’ll get them, but we will ensure we’re in the right place and ready to apply. We want to be in a position to be the national quantum flagship startup in Switzerland in a few years. There’s a lot of public money not only to support academic research but also to drive the commercialization of research outcomes. 

What Advice Would You Give Fellow Deep Tech Founders?

As we just spun out of ETH Zurich, my entrepreneurial journey has been rather short, and the learning curve extremely steep, but knock on wood, things are falling into place rather well. I’d tell other entrepreneurs to believe in themselves, not undersell their potential, and tell their story in a really good and unique way. People want to hear and be part of great stories, and investors like to invest in cool startups. Talk to people about your big vision, and you’ll be surprised by how much support you’ll receive.