Ultra-efficient optical circuits to de-bottleneck data centers and telecom networks.
From the confines of a quantum computer, to data centers, to nondescript cables spanning our oceans or threaded beneath our city streets, optical fiber enables instant and profound connectivity.
The connections between our most fundamental technologies rely on a device to convert signals between electricity and light waves at high speeds: the electro-optic modulator.
Electro-optic modulators made with lithium niobate (LN) are the most common due to LN’s long-known ability to efficiently convert between electrical and optical domains. However, LN has remained difficult to fabricate on the chip scale using microfabrication processes, which has left electro-optic modulators in bulky, discrete, expensive forms that cannot scale, integrate with CMOS electronics, or achieve certain performance metrics. Photonics platforms based on other materials do provide on-chip integration, but come with performance trade-offs due to non-ideal material properties.
As communication demands continue to evolve, so too does the need to create smaller, more accessible modulators that can democratize the power of LN so that it is not just the technology of major telecommunications installations.
HyperLight has done just that.
The team of Mian Zhang, Cheng Wang, and Marko Loncar have created integrated, chip-scale LN modulators. It is an immense breakthrough on an immensely small scale. Through work out of the Laboratory for Nanoscale Optics at Harvard University, the trio discovered a method of fabricating thin LN film modulators with extremely low signal loss. This is a foundational technology, one that will enable future communication networks to operate at higher speeds and lower power, everywhere.
“Imagine,” Zhang begins, “that the long-distance optical fiber cables that run along the ocean floor are communication interstates. High capacity. High speed. Now imagine if we made all the connections between data centers, industries, offices, and homes equally capable. It would be like turning all our back roads into highways. Exponentially more capacity and speed. We could get everything, and get more of it, faster while consuming significantly less energy.”
Manipulating LN at the nanoscale is not easy. The superior qualities of lithium niobate crystals are well known in the photonics industry, but etching the material at the scale needed for low-power, chip-size applications has vexed a generation of physicists and materials scientists. LN has a tendency to stick to itself and the manufacturing substrate. It is a high-risk material problem, as Zhang notes. Failure is likely.
For Zhang, the prospect of building something—creating new and real technology, from bottom up—brought him from the labs of an undergraduate physics department, to a doctoral lab experimenting with silicon photonics, and his postdoctoral research in applied physics at Harvard.
It was at Harvard that he joined the lab of Marko Loncar, a professor renowned for his work with difficult-to-fabricate optical materials like diamond. In Loncar’s lab, Zhang was introduced to Cheng Wang, a PhD student who had just successfully produced LN nanowaveguides that proved the potential of LN’s superior efficiency. The production of these nanowaveguides also proved the lab’s capability to manufacture structures with even greater efficiency and performance.
Zhang, Wang, and Loncar set out to reduce the loss factor of these waveguides by a factor of 10—from losing 50% of light per centimeter of propagation, to losing the same amount over 10cm. It was a goal, if met, that would upend the entire photonics landscape. The experimental results astounded even the team—they reduced the loss factor by a factor of 100. They produced a waveguide in which light could propagate with negligible loss for over one meter. They used these low loss waveguides to make the highest performance electro-optic modulator, showcased in an article in the journal Nature.
The team then realized that the integrated optical modulator devices made using their ultralow-loss chips could meet the growing market demand for ultrahigh-performance, yet cost-effective optical solutions. HyperLight was born.
When Charles Kao, the Nobel Laureate and pioneer of optical fiber communication, made his most significant breakthroughs in the 1960s and 1970s, he could not have predicted how ubiquitous broadband would shape today’s society and global economy. Low-loss optical fiber changed the way the world connects. Zhang, Loncar, and Wang saw this foundational technology as something to be improved, something to be evolved to meet the demands of the next half-century. They’ve engineered that improvement. Their ultralow-loss chips, and the techniques to harness the true potential of lithium niobate, will help us exploit tomorrow’s data, and its connections, with unprecedented levels of speed and efficiency.