A team of engineers has built a hair-thin optical chip that can precisely control laser light using far less power, offering a practical path toward much larger quantum computers. The device is made in standard chip factories, making it cheap and scalable.
A new optical chip, nearly 100 times thinner than a human hair, could help turn today’s lab-scale quantum computers into powerful machines with thousands or even millions of quantum bits.
Researchers report that their tiny device, published in Nature Communications, can precisely reshape laser light while using a fraction of the power of today’s bulky hardware. Just as important, it is made with the same mass-production technology used for everyday computer chips, making it practical to build in huge numbers.
The work, led by Jake Freedman, an incoming doctoral student in the University of Colorado Boulder’s Department of Electrical, Computer & Energy Engineering, Matt Eichenfield, a professor and Karl Gustafson Endowed Chair in Quantum Engineering, and collaborators from Sandia National Laboratories, including co-senior author Nils Otterstrom, tackles one of the biggest bottlenecks in scaling up quantum computers: controlling light.
Many of the most promising quantum computers store information in individual atoms or ions held in place by electromagnetic fields. To operate these qubits, researchers must precisely address each atom with laser beams, using tiny shifts in laser frequency to encode and manipulate quantum information.
In essence, scientists need to be able to reliably “talk” to each atom without disturbing its neighbors.
Freedman explained that making slightly shifted copies of a laser is central to that task.
“Creating new copies of a laser with very exact differences in frequency is one of the most important tools for working with atom- and ion-based quantum computers,” he said in a news release. “But to do that at scale, you need technology that can efficiently generate those new frequencies.”
Right now, those frequency shifts are usually produced by large, table-top electro-optic modulators that sit on optical benches and consume a lot of microwave power. They work well for experiments with small numbers of qubits, but they are not designed for systems that might eventually need tens or hundreds of thousands of separate optical channels.
Eichenfield put the challenge bluntly.
“You’re not going to build a quantum computer with 100,000 bulk electro-optic modulators sitting in a warehouse full of optical tables,” he said in the news release. “You need some much more scalable ways to manufacture them that don’t have to be hand-assembled and with long optical paths. While you’re at it, if you can make them all fit on a few small microchips and produce 100 times less heat, you’re much more likely to make it work.”
The team’s new device is a step in that direction. It is an optical phase modulator, a component that changes the phase of light — essentially, how the peaks and valleys of a light wave line up over time. By controlling phase at very high speeds, the chip can generate new, precisely spaced frequencies from a single laser.
To do this, the device uses microwave-frequency mechanical vibrations that oscillate billions of times per second. These ultra-fast vibrations interact with the light traveling through the chip, imprinting controlled changes on the laser’s phase. The result is a set of new laser frequencies with high stability and efficiency, which are crucial for quantum computing, quantum sensing and quantum networking.
According to the researchers, their phase modulator can generate these new frequencies while consuming about 80 times less microwave power than many commercial devices. Lower power use means less heat, which in turn allows many more channels to be packed closely together, potentially on a single chip.
That combination of efficiency, compactness and precision is what makes the technology promising for future quantum machines that may need to coordinate an intricate “dance” of thousands of atoms at once.
Equally significant is how the device is made. Instead of relying on custom, hand-built components, the team fabricated the modulators entirely in a standard semiconductor manufacturing facility, often called a “fab.”
“CMOS fabrication is the most scalable technology humans have ever invented,” Eichenfield added, referring to the process used to build modern microprocessors and memory chips. “Every microelectronic chip in every cell phone or computer has billions of essentially identical transistors on it. So, by using CMOS fabrication, in the future, we can produce thousands or even millions of identical versions of our photonic devices, which is exactly what quantum computing will need.”
The work takes devices that were once expensive, power-hungry and bulky and makes them more efficient and compact, according to Otterstorm. He described the shift as a turning point for light-based technologies.
“We’re helping to push optics into its own ‘transistor revolution’, moving away from the optical equivalent of vacuum tubes and towards scalable integrated photonic technologies,” Otterstorm added.
In the long run, that kind of “transistor revolution” for optics could do for quantum hardware what integrated circuits did for classical computing in the 20th century: shrink room-sized systems down to chips and make them cheap and reliable enough to deploy widely.
The team is already working on the next step. They are developing fully integrated photonic circuits that combine several key functions — frequency generation, filtering and pulse shaping — onto the same chip. That would move them closer to a complete, ready-to-use optical control platform for quantum computers.
Next, the researchers plan to partner with quantum computing companies to test versions of these chips inside state-of-the-art trapped-ion and trapped-neutral-atom systems. Those collaborations will help show how the new modulators perform under real-world conditions and what refinements are needed.
Freedman sees the device as a crucial milestone on the road to practical, large-scale quantum machines.
“This device is one of the final pieces of the puzzle,” he added. “We’re getting close to a truly scalable photonic platform capable of controlling very large numbers of qubits.”
Source: University of Colorado Boulder