Imagine building a tiny, incredibly efficient light circuit – one so good it rivals the performance of the best optical fibers. That's precisely what researchers have achieved, and it's poised to revolutionize fields from quantum computing to medical imaging. But here's the kicker: they did it using a material that's been around for a while but has been notoriously difficult to work with in this way: germano-silicate.
This breakthrough, detailed in a recent Nature article (https://www.nature.com/articles/s41586-025-09889-w), showcases an ultralow-loss photonic integrated circuit (PIC) platform. The secret? Germano-silicate, a material already celebrated for its exceptional performance in optical fibers, is now being manufactured using a process fully compatible with standard CMOS foundries – the same facilities that churn out your computer chips. This means mass production and lower costs are potentially within reach.
Image Credit: Sergey Saulyak/Shutterstock.com
So, why is this such a big deal? Let's dive deeper.
The Challenge: Short Wavelengths, Big Losses
At shorter wavelengths (think 400–1,100 nm, the violet to near-infrared range), light doesn't travel as smoothly. Two main culprits are to blame. First, there's surface Rayleigh scattering. Imagine shining a flashlight on a bumpy road – the light scatters in all directions. A similar thing happens when light encounters tiny imperfections (surface roughness) on a waveguide. The shorter the wavelength, the more pronounced this scattering becomes. Second, absorption losses increase. Photons at these wavelengths have enough energy to get absorbed by the material itself, especially as you approach what's known as the Urbach tail – a region where materials start absorbing light more readily.
Despite these challenges, many cutting-edge technologies depend on these shorter wavelengths. Think of highly precise optical clocks, the mind-bending world of quantum computing, advanced bioimaging techniques, compact lidar systems (used in self-driving cars, for example - https://www.azooptics.com/Article.aspx?ArticleID=143), and fundamental atomic physics research. These applications need light to travel with minimal loss.
Silica (pure glass) and germano-silicate (silica doped with germanium) have long been the go-to materials for short-wavelength optical fibers because of their remarkably low material absorption. But here's where it gets controversial... While silica works well in fibers, creating planar (flat) integrated circuits with it has been tricky. Pure silica often requires suspended structures, which are delicate and hard to manufacture. Germano-silicate, while promising, lacked a well-established, scalable fabrication process – until now.
The Solution: A CMOS-Compatible Germano-Silicate Platform
The key to this breakthrough lies in a carefully designed fabrication process that's fully compatible with CMOS manufacturing. This is a game-changer because it opens the door to mass production using existing infrastructure.
The process begins by depositing a 4-μm-thick layer of germano-silica (containing 25 mol% GeO2, giving it about a 2% refractive index contrast - https://www.azooptics.com/Article.aspx?ArticleID=28) onto a 15-μm-thick layer of thermal oxide (an insulating layer) on a silicon wafer. This deposition is done using plasma-enhanced chemical vapor deposition (PECVD) at a relatively low temperature of around 270°C. This low temperature is critical because it sets a limit (an “anneal-free thermal budget”) on how much heat the material can withstand during the entire process. Why is that important? Because high temperatures can degrade the material's performance.
Next, the researchers pattern ridge waveguides (the “wires” that guide the light) in the germano-silica layer. They use a combination of ruthenium (Ru) and silica hard masks, deep-ultraviolet (DUV) lithography (a precise way of etching patterns), and inductively coupled plasma (ICP) etching. The ruthenium mask is particularly important. It provides the high etch selectivity needed for the deep etching of the Ge:silica material using fluorine-based chemistry.
And this is the part most people miss... To minimize scattering losses caused by surface roughness and achieve ultrahigh Q factors (a measure of how well a resonator stores energy), the wafer undergoes a furnace annealing step. This heat treatment causes the germano-silica waveguide sidewalls to reflow (smooth out), effectively polishing away etch-induced roughness. The underlying thermal oxide layer remains unaffected during this process.
Finally, an optional upper cladding layer can be added after annealing. The researchers tested two cladding options: a 14-μm-thick phosphorus-doped silica layer for full acoustic confinement (more on that later), and a higher-quality ICP-PECVD silica cladding (more than 6 μm thick) to better protect the devices from long-term exposure to the atmosphere.
The use of DUV-stepper lithography throughout the process ensures high-precision patterning, which is essential for accurately controlling the dispersion (how light of different colors spreads out) and overall device performance.
The Results: Record-Breaking Performance
The germano-silicate PICs demonstrated record-low waveguide propagation losses across the violet to telecom bands. Resonator Q factors exceeded 180 million across this broad spectrum, reaching a peak of 463 million at 1,064 nm. This translates to a waveguide loss of just 0.08 dB per meter – a figure comparable to the very first low-loss optical fibers produced back in 1970!
More importantly, this platform overcomes the short-wavelength limitation. It achieves a loss of only 0.49 dB per meter at 458 nm, a significant 13-dB improvement over previous records in the visible and short-NIR ranges. This is a major step forward.
But the benefits extend beyond just low loss. The platform's material and structural properties allow for advanced functionality. The DUV-stepper-defined waveguides enable precise dispersion engineering, which is crucial for generating soliton microcombs (tiny optical frequency combs with a wide range of applications). The material also facilitates acoustic mode confinement, which was confirmed by characterizing the stimulated Brillouin scattering (SBS) gain spectrum.
Integrated germano-silicate resonators were used to create a high-coherence Brillouin laser, exhibiting a lasing frequency shift of 9.68 GHz. This is lower than the typical 10.9 GHz observed in standard silica resonators, demonstrating the improved acoustic confinement. This combination of ultralow optical loss and engineered acoustic confinement enables low-noise Brillouin lasers, which are essential for advanced gyroscopes and integrated microwave photonics (https://www.azooptics.com/Article.aspx?ArticleID=2687).
Furthermore, the platform's large mode area (LMA) enhances hybrid-integrated low-noise lasers. Self-injection locking (SIL) of semiconductor diode lasers with ultrahigh-Q germano-silicate microresonators significantly reduces frequency noise. The large mode area, calculated at 28.06 μm2 for Ge-silica compared to 7.71 μm2 and 1.33 μm2 for thin and thick Si3N4, respectively, greatly suppresses thermal refractive noise (TRN). As a result, a commercial DFB laser coupled with a Ge-silica resonator achieved a Hz-level fundamental linewidth, corresponding to a 46-dB noise reduction relative to the free-running laser. Extending this into the visible spectrum, SIL of commercial Fabry–Pérot diode lasers resulted in fundamental linewidths of 15 Hz at 632 nm, 12 Hz at 512 nm, and 90 Hz at 444 nm. These incredibly narrow linewidths are essential for precision measurements and sensing.
The Implications: A New Era for Integrated Photonics
In conclusion, this new germano-silicate ultralow-loss platform represents a significant leap forward in integrated photonics, achieving a >10 dB improvement in quality factor in both the violet wavelength range and for anneal-free processing. What does this mean? Better performance and easier manufacturing.
The platform also offers readily engineered dispersion, acoustic mode confinement, and thermal stability, as demonstrated by soliton microcomb generation, stimulated Brillouin lasing, and low-frequency-noise self-injection locking. Crucially, ultralow losses are achieved without post-processing thermal annealing in the telecom band, offering a 10-fold reduction in anneal-free waveguide loss over previous records and enabling easier heterogeneous integration (combining different materials and components on the same chip). The success of this germano-silicate platform promises to bring fiber-like performance to integrated circuits and unlock a new wave of applications, from ultra-precise optical clocks to highly sensitive quantum sensors.
This advancement raises a critical question: Could this germano-silicate platform eventually replace silicon nitride as the dominant material for high-performance PICs? While silicon nitride has its strengths, the superior loss performance and CMOS compatibility of this new platform could give it a significant edge in many applications. Or, perhaps, will there be a synergistic relationship between the two materials, with each being used for specific applications where they excel?
What do you think? Share your thoughts in the comments below!
Download the PDF of this page
Source:
Journal Reference
Chen H. J., Colburn K., et al. (2026). Towards fibre-like loss for photonic integration from violet to near-infrared. Nature 649, 338–344. DOI: 10.1038/s41586-025-09889-w, https://www.nature.com/articles/s41586-025-09889-w
Related Stories
- Menhir Photonics Presents An Improved Version Of Its Menhir-1550 Ultralow-Noise Femtosecond Laser (https://www.azooptics.com/News.aspx?newsID=27428)
- Major OEMs Place Preliminary Orders for Zecotek Photonics’ Lutetium Fine Silicate Scintillation Crystals (https://www.azooptics.com/News.aspx?newsID=22412)
- Zero- to Ultralow-Field NMR Spectroscopy to Observe Reactions in Metal Containers (https://www.azooptics.com/News.aspx?newsID=25152)
