Quantum computers could one day outperform today's supercomputers, and building them with particles of light — photons — as their fundamental units, or qubits, could make them even faster and more energy-efficient. However, fabricating the chips that allow photons to function as qubits at scale has been a major hurdle. Engineers at Purdue University have now reported progress on several of these challenges, according to a university news release.

A qubit is the quantum equivalent of a classical computer's 'bit,' which stores information as either a 0 or a 1. A qubit can exist as both 0 and 1 simultaneously through a phenomenon called superposition, vastly increasing the amount of information it can process. Photons, the smallest particles of light, offer several advantages for qubits over other quantum particles. They can travel through circuit gates at the speed of light, minimising delays, and are less easily disturbed by heat or other environmental changes, allowing them to operate stably at room temperature. Other qubit types require extreme cold, consuming more energy.

Despite these benefits, quantum photonic integrated circuits currently work only with a small number of qubits. Practical quantum computers would need hundreds or thousands of qubits, a scale that remains out of reach. "Single photons are the building blocks of quantum photonics. The biggest challenges in scaling up a quantum photonic integrated circuit have to do with creating good single-photon sources," said Vladimir Shalaev, the Bob and Anne Burnett Distinguished Professor in Electrical and Computer Engineering at Purdue, as quoted in the release.

Shalaev and Alexandra Boltasseva, the Ron and Dotty Garvin Tonjes Distinguished Professor in Electrical and Computer Engineering, lead the work at Purdue's Elmore Family School of Electrical and Computer Engineering, the Birck Nanotechnology Center, and the Purdue Quantum Science and Engineering Institute. Their past discoveries on manipulating light at the nanoscale have been foundational to photonics and optics. Both have been repeatedly named to the Clarivate Highly Cited Researchers list, placing them in the top 1% of citations in the Web of Science. Boltasseva was also featured on last year's Quantum 100 list as part of the UN-designated International Year of Quantum Science and Technology.

A more reliable way to produce single photons

One difficulty in building quantum photonic integrated circuits is generating exactly one photon at a time to serve as a qubit. In 2021, Shalaev and Boltasseva published a study in Science Advances demonstrating a method to emit single photons using silicon nitride, a material that has become the industry standard for quantum photonic platforms. The conventional process, called spontaneous parametric down-conversion, produces single photons inconsistently and with low efficiency, the release states.

The Purdue team's method emits exactly one photon at room temperature from dozens of locations on a silicon nitride platform each time a laser beam passes through it. The photon sources are modifications to the atomic structure — 'defects' — created within the platform, which limit light emission to a single photon at a time. In a study presented at the 2022 Conference on Lasers and Electro-Optics, the researchers showed they could control the placement and number of these defects by embedding them in nanosized pillars on the platform, potentially allowing more uniform manufacturing for use in circuits.

A single-photon switch

Another challenge is that single photons do not readily interact with each other, making them difficult to control on a chip using switches. In a conventional computer, transistors toggle bits between 0 and 1. Similarly, switches in a quantum photonic integrated circuit would enable photons to act as qubits. Last year, the team reported in Nature Nanotechnology that they had built a switch capable of operating at the level of single photons, according to the release.

The proof-of-concept switch adapts a mechanism called optical nonlinearity, which the industry uses to modulate light with another high-intensity light beam for applications such as fibre-optic communications. This mechanism allows a 'control' light beam to interact with a 'probe' beam, but typically at power levels too high for single photons. With the Purdue method, a single photon from the control beam strikes silicon, creating an electron that triggers an avalanche of electrons. This avalanche alters how light passes through the material, changing its refractive index. By changing silicon's refractive index, the probe beam passes through the device only when a single photon starts the electron avalanche, effectively switching the beam on and off.

A tool for predicting quantum decoherence

Even though photon-based qubits are less affected by their surroundings, a single disturbance — called decoherence — can still cause qubits to fall out of superposition or entanglement, creating errors. Purdue electrical and computer engineering professor Alexander Kildishev and graduate student Pranshu Maan led a team that included Shalaev and Boltasseva in demonstrating an approach that could prevent decoherence by anticipating it. The findings were published in Nature Communications this year.

"Decoherence is seemingly random, but in fact there is a certain pattern that can be recognized. This study takes an important step in identifying such a pattern using machine learning," Boltasseva said in the release.

The team found through statistical-physics analysis that the way decoherence unfolds over time follows a pattern. They used machine learning to validate that these patterns contain information that could be predicted. They built a framework that learns and could anticipate the dynamics of a quantum system, enabling proactive control of decoherence rather than only responding after it occurs.

With this framework, the researchers demonstrated a way to predict and reduce one of the major causes of decoherence, called spectral diffusion. This occurs when an emitted photon shifts across the electromagnetic spectrum due to natural disorder in the surrounding environment, changing the photon's frequency and energy level. Because superposition requires all photons in a quantum system to be indistinguishable, a change in a single photon's energy level can break superposition. The researchers discovered that spectral shifts retain a memory of how a photon's frequency has wandered over time, revealing a predictable temporal pattern. The team created a predictive model incorporating a memory function, which they taught using information from single quantum emitters to forecast upcoming spectral changes. By applying these forecasts, they demonstrated that a proactive control method could diminish spectral shifts up to 20 times in the examined systems.

"You can't get rid of decoherence — it's the law of nature that everything fluctuates. But if you recognize the pattern of decoherence and learn how to control it, then you can minimize it," Shalaev said. "Let's say you need a hundred qubits to run a quantum algorithm, but typically far more than that, say 10,000 qubits, to deal with errors caused by decoherence. If I minimize those errors somehow, then the number of qubits needed to run my algorithm would also be minimized."

In principle, this approach could lay the foundation for anticipating decoherence across multiple quantum systems, which is needed for scalable quantum architecture. Those predictions could then be combined with active tuning to keep different emitters more closely aligned, improving the indistinguishability and synchronisation of the photons they generate.

The team next plans to use tools from statistical physics to extract hidden properties of a quantum system's fluctuating environment from its measured dynamics, incorporating these insights into machine-learning models to improve decoherence prediction. They will also investigate whether the models can identify the microscopic processes responsible for spectral shifts, with the long-term goal of predicting and suppressing decoherence before it significantly disrupts the quantum state.

Shalaev and Boltasseva's work related to quantum photonic integrated circuits is based on multiple technologies disclosed to the Purdue Innovates Office of Technology Commercialization, which has applied for patents. The research on the single-photon emitter and the decoherence-prediction algorithm was supported by the U.S. Department of Energy's Quantum Science Center. The algorithm's development also received support from the Purdue-led Quantum Photonic Integrated Design Center, an Energy Frontier Research Center funded by DOE's Office of Science. The work on the single-photon switch was funded by the U.S. Army Research Laboratory.