Un nuevo método para la generación en chip de fotones individuales

Moody Lab ha creado un nuevo enfoque para generar fotones individuales en un chip.

El rumor que rodea el futuro de la tecnología cuántica continúa intensificándose a medida que los investigadores se esfuerzan por aprovechar el potencial de las partículas cuánticas superpuestas, entrelazadas y túneles. Estas partículas tienen la capacidad única de existir en dos estados simultáneamente, lo que podría mejorar en gran medida la potencia y la eficiencia en muchas aplicaciones.

Según Kamyar Parto, estudiante de doctorado en UC Santa Barbara y coautor principal de un artículo publicado en Nano-letras, el estado actual de los dispositivos cuánticos es «más o menos donde estaba la computadora en la década de 1950», o al comienzo de su desarrollo. Parto trabaja en el laboratorio de Galan Moody, un reconocido experto en fotónica cuántica y profesor asistente de ingeniería eléctrica e informática. El artículo detalla un avance significativo en el campo: la creación de una «fábrica» ​​en el chip para generar un flujo constante y rápido de fotones individuales, que son cruciales para el avance de las tecnologías cuánticas basadas en la fotónica.

En las primeras etapas del desarrollo de la computadora, explicó Parto, “los investigadores acababan de hacer el transistor y tenían ideas sobre cómo hacer un interruptor digital, pero la plataforma era bastante débil. Diferentes grupos desarrollaron diferentes plataformas y, finalmente, todos convergieron en CMOS (semiconductor complementario de óxido de metal). Luego tuvimos la gran explosión en torno a los semiconductores.

“La tecnología cuántica está en un lugar similar: tenemos la idea y una idea de lo que podríamos hacer con ella, y hay muchas plataformas competidoras, pero aún no hay un ganador claro”, continuó. “Tienes qubits superconductores, qubits de espín de silicio, qubits de espín electrostático y computadoras cuánticas basadas en trampas de iones. Microsoft está tratando de crear qubits topológicamente protegidos, y en Moody Lab estamos trabajando en fotónica cuántica.

Parto predice que la plataforma ganadora será una combinación de diferentes plataformas, dado que cada una es poderosa pero también tiene limitaciones. «Por ejemplo, es muy fácil transferir información usando fotónica cuántica porque a la luz le gusta moverse», dijo. “Sin embargo, un qubit giratorio hace que sea más fácil almacenar información y hacer ‘cosas’ locales, pero no puedes mover esos datos. Entonces, ¿por qué no tratamos de usar la fotónica para transferir datos desde la plataforma que los almacena mejor y luego los transformamos nuevamente a otro formato una vez que están allí? »

Los qubits, esos controladores de tecnologías cuánticas de comportamiento extraño, son, por supuesto, diferentes de los bits clásicos, que solo pueden existir en un solo estado de cero o uno. Los qubits pueden ser uno y cero simultáneamente. En el campo de la fotónica, dijo Parto, sólo una[{» attribute=»»>photon can be made both to exist (state one) and not to exist (state zero).

That is because a single photon constitutes what is called a two-level system, meaning that it can exist in a zero state, a one state, or any combination, such as 50% one and 50% zero, or maybe 80% one and 20% zero. This can be done routinely in the Moody group. The challenge is to generate and collect single photons with very high efficiency, such as by routing them on a chip using waveguides. Waveguides do exactly what their name suggests, guiding the light where it needs to go, much as wires guide electricity.

Parto explained: “If we put these single photons into many different waveguides — a thousand single photons on each waveguide — and we sort of choreograph how the photons travel along the waveguides on the chip, we can do a quantum computation.”

While it is relatively simple to use waveguides to route photons on chip, isolating a single photon is not easy, and setting up a system that produces billions of them rapidly and efficiently is much harder. The new paper describes a technique that employs a peculiar phenomenon to generate single photons with an efficiency that is much greater than has been achieved previously.

“The work is about amplifying the generation of these single photons so that they become useful to actual applications,” Parto said. “The breakthrough described in this paper is that we can now generate the single photons reliably at room temperature in a way that lends itself to (the mass-production process of) CMOS.”

There are various ways to go about generating single photons, but Parto and his colleagues are doing it by using defects in certain two-dimensional (2D) semiconductor materials, which are only one atom thick, essentially removing a bit of the material to create a defect.

“If you shine light (generated by a laser) onto the right kind of defect, the material will respond by emitting single photons,” Parto said, adding, “The defect in the material acts as what is called a rate-limiting state, which allows it to behave like a factory for pushing out single photons, one at a time.” One photon might be produced as often as every three to five nanoseconds, but the researchers aren’t yet sure of the rate, and Parto, who earned his Ph.D. on the topic of engineering such defects, says that the current rate could be much slower.

A big advantage of 2D materials is that they lend themselves to having defects engineered into them at specific locations. Further, Parto said, “The materials are so thin that you can pick them up and put them on any other material without being constrained by the lattice geometry of a 3D crystal material. That makes the 2D material very easy to integrate, a capability we show in this paper.”

To make a useful device, the defect on the 2D material must be placed in the waveguides with extreme precision. “There is one point on the material that produces light from a defect,” Parto noted, “and we need to get that single photon into a waveguide.”

Researchers try to do that in a couple of ways, for instance, by putting the material on the waveguide and then looking for an existing single defect, but even if the defect is precisely aligned and in exactly the right position, the extraction efficiency will be only 20% to 30%. That is because the single defect can emit only at one specific rate, and some of the light is emitted at oblique angles, rather than directly along the path to the waveguide. The theoretical upper limit of that design is only 40%, but making a useful device for quantum-information applications requires 99.99% extraction efficiency.

“The light from a defect inherently shines everywhere, but we prefer that it shine into these waveguides,” Parto explained. “We have two choices. If you put waveguides on top of the defect, maybe ten to fifteen percent of the light would go into the waveguides. That’s not enough. But there is a physics phenomenon, called the Purcell effect, that we can utilize to boost this efficiency and direct more of the light into the waveguide. You do that by placing the defect inside an optical cavity — in our case, it’s in the shape of a micro-ring resonator, which is one of the only cavities that allows you to couple light into and out of a waveguide.

“If the cavity is small enough,” he added, “it will squeeze out the vacuum fluctuations of the electromagnetic field, and those fluctuations are what cause the spontaneous emission of photons from the defect into a mode of light. By squeezing that quantum fluctuation into a cavity of finite volume, the fluctuation over the defect is increased, causing it to emit light preferentially into the ring, where it accelerates and becomes brighter, thus increasing the extraction efficiency.”

In experiments using the micro-ring resonator that were done for this paper, the team achieved an extraction efficiency of 46%, which is an order-of-magnitude increase over prior reports.

“We’re really encouraged by these results because single-photon emitters in 2D materials address some of the outstanding challenges facing other materials in terms of scalability and manufacturability,” said Moody. “In the near term, we’ll explore using them for a few different applications in quantum communications, but in the long term, our goal is to continue to develop this platform for quantum computing and networking.”

To do that, the group needs to improve their efficiency to better than 99%, and achieving that will require higher-quality nitride resonator rings. “To enhance efficiency, you need to smooth out the ring when you carve it out of the silicon nitride film,” Parto explained. “However, if the material itself is not fully crystalline, even if you try to smooth it at the atomic level, the surfaces could still look rough and sponge-like, causing the light to scatter off of them.”

While some groups achieve the highest-quality nitride by purchasing it from companies that grow it perfectly, Parto explained, “We have to grow it ourselves, because we have to put the defect under the material, and also, we’re using a special type of silicon nitride that minimizes the background light for single-photon applications, and the companies don’t do that.”

Parto can grow his nitrides in a plasma-enhanced chemical vapor deposition oven in the cleanroom at UCSB, but because it is a heavily used shared facility, he is not able to customize some settings that would allow him to grow material of sufficient quality. The plan, he says, is to use these results to apply for new grants that would make it possible “to get our own tools and hire students to do this work.”

Reference: “Cavity-Enhanced 2D Material Quantum Emitters Deterministically Integrated with Silicon Nitride Microresonators” by K. Parto, S. I. Azzam, N. Lewis, S. D. Patel, S. Umezawa, K. Watanabe, T. Taniguchi and G. Moody, 1 November 2022, Nano Letters.
DOI: 10.1021/acs.nanolett.2c03151