The path to a useful quantum computer
Erik Hosler is the lead for the Process Exploration for Photonics Department at PsiQuantum. This group is responsible for looking at potential materials, processes, and architectures that will optimize component performance of their silicon photonic quantum computer. These are “high-risk, high-reward” activities that could have a dramatic impact on the performance of the machine. Hosler's role is to ensure that the team has the resources available to explore the best possible processes, while keeping the focus on the goal of building the world’s first useful quantum computer. He will be giving a plenary talk on this subject at SPIE Advanced Lithography + Patterning.
What are some of the projects in this department that you’re most excited about?
I enjoy trying to be clever, and part of our mission is to figure out how to drive process perfection without breaking the bank or the fab. Line-edge roughness (LER) and critical dimension uniformity (CDU) are two parameters that dictate the performance of a large set of photonic and superconducting devices, so the closer we can drive these characteristics to the atomic limit, the better the machine functions.
Specifically, backporting leading-edge CMOS capabilities to our process is one of the ways that we can build off the momentum of the entire high-volume manufacturing semiconductor industry to deliver photonic devices that are truly revolutionary. Figuring out how to squeeze seemingly impossible performance in a creative, novel way is really what gets me excited every day!
How would you define a truly useful quantum computer?
In the simplest sense, a useful quantum computer is one in which the value of the computations it performs exceeds the cost of building and operating the machine. To that end, a truly useful quantum computer is one that is fault-tolerant and can tackle whatever problem you throw at it without the need for unique circuitry for each problem (i.e., a general-purpose machine).
For this, you need on the order of a million physical qubits. At PsiQuantum we are directly going for this bold machine design by leveraging a mature, 300mm silicon photonics process through our partnership with GlobalFoundries. Between GlobalFoundries’ product portfolio and our internal expertise on photonics and quantum architecture, we are on track to deliver something truly useful.
How do you think quantum computing will revolutionize our daily lives?
Like the introduction of conventional computers, quantum computers will change the way we look at the world. We won’t have quantum computers in our pockets like our mobiles, but we’ll have access to these machines to do cutting-edge work across many industries. Large-scale, fault-tolerant quantum computers are anticipated to unlock the solutions to otherwise intractable problems and enable extraordinary advances across a broad range of applications including climate, healthcare, life sciences, energy and beyond. Whether it’s improving carbon capture catalysts, optimizing the energy grid, modelling the chemistries of lifesaving drugs or new battery materials, quantum computers are key to solving many of the world’s most demanding challenges that will forever be beyond the capabilities of any conventional computer.
To give an analogy: EUV machines are about the least accessible, most exclusive, esoteric technology that exists. There are very few of them, they are very expensive, and very hard to use. And yet, in any given coffee shop, there are probably hundreds of chips which would not have been made without an EUV machine, and every day we use technology which is built on a foundation of incredibly advanced lithography. In the same way, you should not expect quantum computers to be user-friendly, cheap, or widely available any time soon. But the products of the otherwise impossible calculations are expected to be ubiquitous in everyday life.
What are some of the roadblocks to creating such a device?
The big roadblocks to quantum computing used to be the basic science — creating, entangling, and measuring qubits. Basically, everyone can do that now, regardless of the underlying platform. Qubits are basically a solved problem. The big challenges now are manufacturability, cooling power, connectivity, and control electronics.
Interestingly, especially for SPIE, these are engineering problems: they are scaling problems; they aren’t really quantum problems. We are seeing all the big players now start to tackle these challenges for their respective qubits. And fortunately for us, photonics seems to have orders-of-magnitude-type advantages which give us confidence that we can overcome these challenges.
We are building a utility-scale, fault-tolerant quantum computer with a silicon photonics-based architecture that enables manufacturing in a conventional silicon chip foundry. The photonic approach taken by PsiQuantum has profound technical advantages at the scale required for error correction to deliver a useful quantum computer. We have partnered with GlobalFoundries to leverage existing semiconductor manufacturing technology to achieve this objective, and our manufacturing presence at GlobalFoundries is an unprecedented economic signal of maturity for a technology that is often viewed as being at the early research phase.